Merging categorical variables
Sometimes it is useful to merge categorical variables, for example, because you want to create a total awareness variable or because you want two categorical variables to be treated as one variable in your tables. You do this by creating a new categorical variable based on the existing categorical variables you want to merge. The new categorical variable will automatically have the same categories as the categorical variables you are merging. If the same category appears in more than one of the variables, it will only appear once in the new variable.
To merge existing categorical variables
1 In the main UNICOM Intelligence Reporter - Survey Tabulation window, select the Define tab.
2 In the Variable List, select the categorical variables you want to merge. Make sure that these are not derived variables.
If you are using the hierarchical view of the data, you also need to make sure that the variables you select are at the same level. The position of the variables you select will define the level of the new variable. For example, if you select a top-level variable, the new variable will be at the top level. Whereas, if you select a variable in the Person loop, the new variable will be at the Person level.
3 On the UNICOM Intelligence Reporter - Survey Tabulation menu, click Variables > Merge.
The New Merged Variable dialog opens: see
New Variable dialog.
5 In the Label box, enter a description for the new variable as you want it to appear on your tables.
6 If necessary, use the features on the Edit tab to define how you want the new variable to appear in your tables.
7 Click OK.
Notes
▪If any of the variables you are merging has an axis expression, it will not be copied to the new variable. For example, suppose you merge two categorical variables from which you have deleted one or two categories. The new variable will include the deleted categories.
▪When you merge variables that contain any built-in special elements, the new variable will not contain any of the built-in special elements. You are most likely to see this when you are working with Quanvert data, which typically contains built-in special elements. For example, if you merge two Quanvert variables, either or both of which contain a built-in mean, the new variable will not contain a mean. Similarly, built-in bases are not copied to the new variable and UNICOM Intelligence Reporter - Survey Tabulation will therefore insert an autobase into the new variable.
See also