Cumulative frequency
In a nutshell
The data in grouped frequency tables can be represented in cumulative frequency graphs. From these graphs, the median, quartiles and percentiles of a data set can be estimated.
Drawing a cumulative frequency graph
Cumulative frequency graphs can be drawn using the information from a grouped frequency table. Follow the steps in the table below:
procedure
1. | Add a cumulative frequency column to a grouped frequency table. |
2. | Calculate cumulative frequency by adding frequencies up to and including current row. |
3. | Draw axes where the y axis is the cumulative frequency and x axis is a given measurement. |
4. | Plot points at the upper bound of each class at their respective cumulative frequency. |
5. | Connect the points with a smooth curve or a straight line depending on what fits the graph best. |
Calculating percentiles
Percentiles can be found from a cumulative frequency graph, where the appropriate frequency is taken and the value is found on the x axis using the curve. This frequency can be found using the equation below where k is the percentile and n is the total frequency. This is only valid for continuous data.
100k(n)
Note: The lower quartile can be considered the 25th percentile, the median 50th and upper quartile 75th.
Example 1
The grouped frequency table below shows the height of Mike's classmates. Draw a cumulative frequency graph to represent this data. Then find the lower quartile, median and upper quartile.
Height (cm) | Frequency |
140<h≤145 | 2 |
145<h≤150 | 3 |
150<h≤155 | 13 |
155<h≤160 | 21 |
160<h≤165 | 5 |
Add a cumulative frequency column and calculate each value:
Height (cm)
| Frequency
| Cumulative frequency
|
140<h≤145
| 2
| 2
|
145<h≤150
| 3
| 2+3=5
|
150<h≤155
| 13
| 5+13=18
|
155<h≤160
| 21
| 18+21=39
|
160<h≤165
| 5
| 39+5=44
|
Draw a cumulative frequency graph. Plot points at the upper bound of each class (145 cm,150 cm,155 cm,160 cm and 165 cm) and connect the points using a smooth line:
Find the y value for the lower quartile, median and upper quartile:
LQMedianUQ=0.25(44)=0.5(44)=0.75(44)=11=22=33
Use the dashed lines to extract the respective x values:
The lower quartile is 152.5 cm, the median 155.5 cm and the upper quartile 158 cm.
Note: The value given at any given percentile is only an estimate as this method assumes an even distribution within each class.