Working With Variables

The variables distributed with a database are known as "standard variables”. The standard variables in databases provided with the SEER*Stat software are formatted for common use. However, you will often need to add or remove groupings for your particular analysis. SEER*Stat dictionaries allow you to create two types of variables. User-defined variables are created using one variable as the template; that is, new groupings are defined based on values from only one variable.  Merged variables are variables created using two or more variables.

For example, to create a variable to stratify results by age (e.g. ages < 65 and ages 65+) you would need to define groupings based on values from one variable, age. In this case, it would be simplest to create a user-defined variable.  However, two variables, cancer site and age, are required to create a variable to stratify results by the following groups: "Lung Cancer for Ages < 65", "Lung Cancer for Ages 65+”,  "Colon Cancer for Ages < 65”, and "Colon Cancer for Ages 65+”. To create a variable with these groupings, you must create a merged variable.

A calculated variable is a variable that is not coded in the database; that is, the field is not on the tumor record. SEER*Stat determines its values based on the values of other variables and system specifications. For example, Age at Prevalence Date is a calculated variable used in Limited-Duration Prevalence sessions. The values of this variable are determined from the prevalence date selected for the analysis and either date of birth or date and age at diagnosis.   User-defined variables based on calculated variables are displayed in a separate folder labeled "Calculated.”

For detailed instructions on working with variables in the dictionary, refer to one of the following topics: