When using choice models, there are special data table rules that you can apply depending on whether you are using a default choice set, subject data with response data, or logistic regression.
If in every trial, you can choose any of the response profiles, you can omit the Profile ID Choices selection under Pick Role Variables in the Response Data section of the Choice launch window. The Choice Model platform then assumes that all choice profiles are available on each run.
If you have subject data in the Response data table, select this table as the Select Data Table under the Subject Data. In this case, a Subject ID column does not need to be specified. In fact, it is not used. It is generally assumed that the subject data repeats consistently in multiple runs for each subject.
Ordinary logistic regression can be performed with the Choice Modeling platform.
Note: The Fit Y by X and Fit Model platforms are more convenient to use than the Choice Modeling platform for logistic regression modeling. This section is used only to demonstrate that the Choice Modeling platform can be used for logistic regression, if desired.
If your data are already in the choice-model format, you might want to use the steps given below for logistic regression analysis. However, three steps are needed:
• Create a trivial Profile data table with a row for each response level.
• Put the explanatory variables into the Response data.
• Specify the Response data table, again, for the Subject data table.
For examples of conducting Logistic Regression using the Choice Platform, see Example of Logistic Regression Using the Choice Platform and Example of Logistic Regression for Matched Case-Control Studies.