Documenting your SPSS data sets

Steve Simon


This page is currently being updated from the earlier version of my website. Sorry that it is not yet fully available.

*Dear Professor Mean

Dear Baffled,

The only things people remember from my training classes are my jokes. The rest is unimportant.

It’s great that you want to take the time and effort to document your data. SPSS can rapidly produce dozens of graphs and hundreds of tables. Without good documentation as part of these graphs and tables, you can easily get lost. A little time spent now in data documentation will save you a lot of time later when you are interpreting your output.

Short explanation

In versions 10 and later of SPSS

Use the DATA VIEW tab to enter your data


More details

There are a lot of important ways in which you can document your data. Start with the variable name itself

Variable name

When documenting your data

Please spend some time to provide descriptive variable names. These names have to be 8 characters or less. It can be a mxiture of numbers and letters

The variable name restrictions on length

After you provide a variable name

middle of the DATA | DEFINE VARIABLES dialog box. There are four button here: TYPE

Format type

Click in theTYPE column to add or change the format type. You will notice a gray button on the right hand side. Click on it to get VARIABLE TYPE dialog box.

This dialog box has information about the type of data that you want to use. The most common data type is **NUMERIC

For data with number codes or count data

Select the STRING options for data that is all letters or a mixture of letters and numbers.When you select this option

In general

If you click on the DATE option

Variable and value labels

Click on the label button to get variable labels. A variable label is a longer description of your data. Variable labels appear in your output and make it easier to follow what is going on. You can use a mixture of upper and lower case here

You can put blanks and special symbols in your variable label. If you are very excited about a variable

You can also specify value labels in this dialog box. Value labels provide informative names for levels in any categorical variable. Leave the value labels blank for continuous data like weight or height. They do make sense

Value labels have to be defined one by one. **Type in the number (or letter) code for your category in VALUE field

Missing value codes

If needed

Be careful about missing values. Values can be missing because the subject dropped out of the study. Perhaps you are looking for chemical concentrations that are sometimes too low for a laboratory to detect. Perhaps a subject refused to respond to a certain question. Perhaps you are asking for something like a spouse’s age that is not applicable for a single person. Make sure you understand why your data is missing and discuss this issue with anyone you are consulting with. The statistical handling of missing values can vary greatly depending on how the value came to be missing.

When you are planning your project

Column format

I usually ignore the COLUMN FORMAT button


Let’s see how to document a column of data that represents marital status. **Marital status is a categorical variable with five codes (1=single

We use numeric codes for this variable

A nice variable label is “Marital Status of the Infant’s Mother”. Notice that we can include an apostrophe here. I also used a mixture of upper and lower case. This is easier to read than all lower case and much easier than all upper case.

The value labels are “Single”; “Married”; “Divorced”; “Widowed”; and “Unknown”. Notice again that I use mixed case. Value labels are appropriate here because this is a categorical variable. For a continuous variable like birth weight



Baffled Bill needs to provide some documentation to SPSS data sets that he is creating. Professor Mean explains that you add documentation by selecting DATA | DEFINE VARIABLE from the SPSS menu or double clicking on the column header. You can then provide information about the variable name

You can find an earlier version of this page on my original website.