Cumulative frequency gives the cut-off value or the phenomenon in which the application will work below a certain fixed limit. This tells the occurence of a phenomenon less than a reference value. This is also refereed by the name Frequency of NON -Excedence. This is also some times referred as ogive.

Cumulative frequency corresponding to a certain value is the sum of all frequencies that adds up diagonally to give the total value of the frequency.

Less than cumulative frequency distribution

Lets consider this distribution

Class |
Frequency |
Cumulative frequency |

-0.5 - 9.5 | 6 | 6 |

9.5 - 19.5 | 22 | 28 |

19.25 - 29.5 | 56 | 84 |

29.5 - 39.5 | 33 | 117 |

39.5 - 49.5 | 5 | 121 |

Total |
121 | |

When less than cumulative frequency distribution is used for graphing, always consider take the upper limit of the class intervals. So the table changes to

Upper limit |
Cumulative frequency |

9.5 | 6 |

19.5 | 28 |

29.5 | 84 |

39.5 | 117 |

49.5 | 121 |

Total |

Hence the cumulative frequency graph will be as follows:

Greater than cumulative frequency distribution

Lets consider this distribution

Class |
Frequency |
Cumulative frequency |

-0.5-9.5 | 6 | 112 |

9.5-19.5 | 22 | 115 |

19.5-29.5 | 56 | 93 |

29.5-39.5 | 33 | 37 |

39.5-49.5 | 5 | 5 |

Total | 121 |

When greater than cumulative frequency distribution is used for graphing, always consider take the lower limit of the class intervals. So the table changes to

Upper limit |
Cumulative frequency |

-0.5 | 112 |

9.5 | 115 |

19.5 | 93 |

29.5 | 37 |

39.5 | 5 |

Total |

Hence the graph will be as follows:

Cumulative frequency is of more than or less than type. In order to construct a less than frequency we add up all the frequencies at or above the given class and for a more than type we add up all the frequencies from or below the class. It is always better to arrange the data classes in ascending order in order to determine the cumulative frequency.

Lets check an example with the following cumulative distribution

Class |
Cumulative frequency |

0 - 12 | 6 |

10 - 20 | 28 |

20 - 30 | 84 |

30 - 40 | 117 |

40 - 50 | 121 |

The relative frequency distribution is as follows

Class |
Cumulative frequency |
Relative cumulative frequency |

0 - 10 | 6 | $\frac{6}{121}$ = 0.049 |

10 - 20 | 28 | $\frac{28}{121}$ = 0.231 |

20 - 30 | 84 | $\frac{84}{121}$ = 0.694 |

30 - 40 | 117 | $\frac{}{117121}$ = 0.967 |

40 - 50 | 121 | $\frac{121}{121}$ = 1 |

The total frequency of all classes that are less than its upper class boundary is called cumulative frequency distribution. It can also be defined as the total frequency of all classes that are greater than its lower class boundary is called cumulative frequency distribution.

Hence, cumulative frequency distribution can be used to determine the number of data that lie either above or below a particular value in a data set.

Given a frequency distribution table, its rather easy to calculate the cumulative frequency.

A tabular arrangement of data showing how many observations lie above, or below, certain values is called cumulative frequency distribution.

Determine the frequency distribution from the given cumulative distribution table.

Class Interval | Cumulative frequency |

15 - 20 | 50 |

20 - 25 | 45 |

25 - 30 | 36 |

30 - 35 | 23 |

35 - 40 | 15 |

40 - 45 | 8 |

We can obtain the frequency of any class by taking the difference between its cumulative frequency and that of the next class.

Class Interval | Frequency |

15 - 20 | 50 - 45 = 5 |

20 - 25 | 45 - 36 = 9 |

25 - 30 | 36 - 23 = 13 |

30 - 35 | 23 - 15 = 8 |

35 - 40 | 15 - 8 = 7 |

40 - 45 | 8 - 0 = 8 |

As seen from the above table the frequency of class 30 - 35 is 8, class 35 - 40 is 7 and so on.

The following statistics problems explain the concept of cumulative frequency distribution.

Construct the cumulative frequency distribution for the data recorded from 35 families regarding number of hours spent watching television per day.

Class (Hours) | Frequency (# of Families) | Less than cumulative frequency | More than cumulative frequency |

0 | 5 | 5 | 35 |

1 | 7 | 12 | 30 |

2 | 5 | 17 | 23 |

3 | 3 | 20 | 18 |

4 | 5 | 25 | 15 |

5 | 5 | 30 | 10 |

6 | 2 | 32 | 5 |

7 | 2 | 34 | 3 |

8 | 1 | 35 | 1 |

In the above table, number of families watching less than or equal to 3 hours of television is obtained by adding up all the frequencies above and at data value 3. So we get the cumulative frequency as 20 (5 + 7 + 5).

Similarly number of families watching more than or equal to 5 hours of television is obtained by adding all the frequencies below and at data value 5. So the more than cumulative frequency here is 10 (5 + 2 + 2 + 1).

If it is mentioned just to determine the cumulative frequency we assume it to be the less than type.

Construct cumulative frequency for the frequency distribution of wages of 50 workers in a factory.

Class (Wages) | Frequency (Number of Workers) | Cumulative frequency |

22 - 24 | 3 | 3 |

25 - 27 | 3 | 6 |

28 - 30 | 6 | 12 |

31 - 33 | 12 | 24 |

34 - 36 | 8 | 32 |

37 - 39 | 6 | 38 |

40 - 42 | 5 | 43 |

43 - 45 | 4 | 47 |

46 - 48 | 2 | 49 |

49 - 51 | 1 | 50 |

From the above table we can very well say that the number of workers earning less than $30 is 12. Such information becomes very useful when a data set is huge.

We can also determine the individual frequency of a class if we are given just the cumulative frequency distribution table as shown in the next example.