Model Fitting: Generalized Linear Models


Variables Tab

You can use the Variables tab to specify the variables for the Generalized Linear Models analysis. The Variables tab is shown in FigureĀ 24.6.

For most response distributions, you only need to specify a single response variable in the Y Variables list. If you specify two numeric variables, the analysis assumes that the variables contain count data for a binomial experiment. The value of the first variable is the number of positive responses (or events). The value of the second variable is the number of trials. In this case, the response distribution is automatically set to binomial.

The dialog box supports multiple explanatory variables. You can include nominal variables in the model by adding them to the Classification variables list. You can include interval variables in the model by adding them to the Quantitative variables list.

When you add an explanatory variable, that main effect is added to the Effects tab. You can add interaction effects and nested effects by using the Effects tab.