Model Fitting: Generalized Linear Models

Create the Offset Variable

As discussed earlier in this example, the quantity $\log (\Variable{months})$ is an offset variable for this model. The following steps use the Variable Transformation Wizard to create this variable. (This wizard is described in further detail in Chapter 32: Variable Transformations.)

  1. Select Analysis →Variable Transformation from the main menu.

    The Variable Transformation Wizard in Figure 24.14 appears.

    Figure 24.14: Selecting a Transformation

    Selecting a Transformation


    The transformation log(Y+a) is highlighted by default. Since this is the desired transformation, you can proceed to the next page of the wizard.

  2. Click Next.

    The wizard displays the page shown in Figure 24.15. Note that the transformation appears in the page’s title bar.

    Figure 24.15: Selecting a Variable and a Parameter

    Selecting a Variable and a Parameter


  3. Select the months variable, and click Set Y.

  4. Click Finish.

    Because there are six observations for which months=0, a warning message appears (Figure 24.16) that informs you that the transformed values for these observations are set to missing values.

    Figure 24.16: A Warning Message

    A Warning Message


  5. Click OK to dismiss the warning message.

    The new variable is named Log_months. It contains six missing values. Observations with missing values for the explanatory variables (including the offset variable) or the response variable are not used in fitting the model.