After a data set is collected, but before it is statistically analyzed, it is usually put in an easily understandable order. Frequency distributions and graphs are commonly used for this purpose. A graph presents, in an obvious manner, the shape or distribution of the data, and is a powerful visual aid in data presentation. A frequency distribution groups data into categories, and lists the number of observations which fall into such groups. In addition, the frequency distribution may contain more information such as a cumulative frequency column, etc.

There are two main reasons for constructing frequency distributions. The first is purely descriptive: the author wants to describe the data set. The second reason is as an aid to simplifying large data sets when statistical computations have to be done by hand. Thus, if a researcher has a large data set, and needs to compute statistics by hand, then she will probably group the data. With frequent use of computers today, more and more researchers are entering data sets into computers which do the calculations. The results of these computations using un-grouped data are more precise than those obtained with data grouped into frequency distributions. Because the former should be preferred over the latter, frequency distributions are infrequently used as a step in statistical calculations.