Usually, numerical variables by default will have a frequency of 1 most of the time making the frequency table impractical for any analysis.
| Dataset | Frequency |
| 1 | 1 |
| 9 | 1 |
| 22 | 1 |
| 24 | 1 |
| 32 | 1 |
| 41 | 1 |
| 44 | 1 |
| 48 | 1 |
| 57 | 1 |
| 66 | 1 |
| 70 | 1 |
| 73 | 1 |
| 75 | 1 |
| 76 | 1 |
| 79 | 1 |
| 82 | 1 |
| 87 | 1 |
| 89 | 1 |
| 95 | 1 |
| 100 | 1 |
| Total | 20 |
When dealing with numerical variables, it makes much more sense to group data into intervals and find the corresponding frequencies of the intervals. This way, we make a summary of the data that allows for a meaningful visual representation.
The correct choice of interval width largely depends on the data we are working with.
simple Interval Width formula = (largest number – smallest number)/no of desired intervals
| Desired intervals | 5 |
| Interval width = roundup(100-1)/5 | 20 |
A number is included in an interval if the number is:
- Greater than the lower bound
- lesser than or equal to the upper bound
| Interval start | Interval end | Frequency | Relative frequency |
| 1 | 21 | 2 | 10% |
| 21 | 41 | 4 | 20% |
| 41 | 61 | 3 | 15% |
| 61 | 81 | 6 | 30% |
| 81 | 101 | 5 | 25% |