It's very crucial to check the shape of distribution of the random variable, which is not easy to examine with the help of frequency diagram or table. Frequency table helps us display tally of numerous data observations on various categories.

In most cases of discrete quantitative data & qualitative data we use all the noted values of the categories. But, when there are huge number of various observations then consecutive observation might be grouped each other for forming the combined categories.

Example 1: For instance (growing coffee beans).

We could create the frequency table as shown below:

We don't have to amalgamate them because we have only 10 categories.

But, when it comes to continuous data, selection of categories are more arbitrary as we normally apply 8 to 12 consecutive intervals that are non-overlapping of same width. Lesser than this might be good for smaller sample size and greater for bigger sample. Note that the intervals should cover the whole observed value range.

Example 2: Survival analysis

We've got the following table:

We need to define the intervals as same as the data. Hence, while the data is defined to the closer values or integers then the intervals need to as mentioned in the above table. On the other hand, intervals should be one decimal place while the data is defined. Again, while the data are rounded to its nearest integers then the intervals are normally 0-39.5, 39.5-79.5 so forth. It's also crucial to sum up sample size with table.