Two particularly valuable features are available from the   Transformations menu: Recode, and Compute. The purpose of a recode is very simple. Imagine a variable that is coded from 1 to 5. Sometimes extreme values are not selected by very many individuals. Thus, it may be desirable to combine individuals who responded with either a 4 or a 5 into a single category such as 4. The recode feature is the way to do this. Another situation that often calls for a recode is when a variable is part of a scale but the scoring needs to be reversed before it can be added to other items to make a total score. This is also accomplished with the recode command. To do a recode, click Transformations > Recode > Into Same Variables... or Into Different Variables... and enter the required information. The choice of recoding into the Same or Different Variables is a question of whether it is desirable to preserve the old data. By doing the recode into a Different variable, the old data can be preserved in case a mistake is made or another recoding procedure is tried.

The Compute... command is also under the Transformations menu. This command allows the researcher to construct an equation for changing the scale of a variable. The main usefulness of the procedure is for remedying the situation where the raw data do not meet statistical assumptions. A transformation using the Compute... command can often bring the data back into conformity with statistical assumptions. Most statistics books have a discussion of the various common types of transformations and their potential benefits.