A frequency table displays counts and percentages for each distinct value found in a variable (normally a categorical variable). Frequency tables are produced using the Frequency or from the of the data editor

From the menu choose to produce frequency tables for one or more variables. The dialog contains the usual variable list as well as several buttons:

• Statistics Displays descriptive statistics in addition to the table.
• Charts Produces optionally a bar-chart, a pie-chart or a histogram).
• Format Options to control the appearance of the table.

The default frequency table looks like one of the two presented below.

This is the table you get if the default options for tables are set to display only and . This table is good for publication and presentation.

This is the table you get if the default option for tables are set to display only and as well as and . In practice, this kind of table is much better while working and analyzing data, as you use to refer to variables, and for instance in recoding operations.

These settings can be modified with .

Let us now have a brief look at the "anatomy" of a frequency table.

The body of the frequency table contains the following summary data for each distinct value found in the variable

• Frequency (count) The number of occurrences of the corresponding code (value).
• Percent Percent occurrence of the code (count divided by total number of observations.
• Valid Percent Percent occurrence of the code: count divided by the total number of valid (=non-missing) observations. [Same as previous percentage if the variable does not have any missing value]
• Cumulative Percent Valid percentage cumulated. Usually only meaningful for ordered variables and only useful if you are interested in the total percentage up to a particular category starting with the first.

Some important terms:

• : used to refer to variables and use them in analysis.
• are used to describe variables more fully [long(er) descriptive text]
• , useful to attach a descriptive label to codes of a (categorical) variable.
• Note that all variables have names, but labels are optional.

The Statistics button lets you request many descriptive statistics, however you should be aware that many of these statistics require interval scaled variables, i.e. are not applicable to e.g. nominal variables.

With the Format button you can control how the frequency table is displayed (sorting frequencies, suppressing the production of all [to produce only statistics or charts] or big tables [too many frequencies]) or how frequency table requests for many variables are handled.

The Charts button lets you request barcharts, pie charts and histograms (all charts are also available from the menu). Again make sure to understand that not all charts are suitable for all kinds of variables, namely histograms are meant for interval scaled variable (as well as the addition of the normal curve); pie and bar charts look quite strange with (too) many categories.