Training

If you have IBM SPSS Modeler and IBM SPSS Statistics installed on your PC, you can use a command that is available in IBM SPSS Statistics to analyze your IBM SPSS Modeler data. Although IBM SPSS Modeler has a wide variety of Graphing, Modeling, and Output nodes, there are certain operations that you may either prefer to do in IBM SPSS Statistics or that IBM SPSS Modeler does not provide (e.g., ordinal regression). IBM SPSS Modeler 13 includes several nodes that are specifically designed to let you run any IBM SPSS Statistics transformation command or procedure. The only requirement is that you need to be licensed to run the command in IBM SPSS Statistics. For example, if you are not licensed to run IBM SPSS Custom Tables in IBM SPSS Statistics, you will not be able to run it in IBM SPSS Modeler either. However, you can execute saved IBM SPSS Statistics syntax files when analyzing new data with IBM SPSS Modeler.

To illustrate, imagine you have a series of Likert scale questions with values ranging from 1 to 3 and you want to count the number of times a respondent answered with a “1,” which means “strongly agree.” The Statistics Count command is a natural for this type of operation. To run this command in IBM SPSS Modeler:

1. Add a Statistics Transform node from the IBM SPSS Statistics palette to the stream
2. Open up the node
3. Click on the Select Dialog button to see a list of Statistics Transformation commands

4. Click on Count Values within Cases to bring up the associated Statistics Dialog box
5. Select the variables you want and set up the Count command the same way you would in IBM SPSS Statistics

6. Click OK to see the completed command (not shown)
7. Click OK to run the command

The new field, "satisfied," is now ready for you to analyze within IBM SPSS Modeler.

*IBM SPSS Statistics and IBM SPSS Modeler, formerly called SPSS Statistics and Clementine®, are part of SPSS Inc.’s Predictive Analytics Software portfolio.