# GLM Auxiliary Regression Model

The Auxiliary Regression Model dialog box specifies the model that is used to test for heteroskedasticity.

- Use predicted values
- Uses a model that consists of a constant term, a term that is linear in the predicted values, a term that is quadratic in the predicted values, and an error term.
- Use univariate model
- Uses the model that is specified on the Model subdialog. An intercept term is included if the specified model does not contain one.
- Custom model
- Uses the model that you explicitly specify.
- Build terms
- Use this choice when you want to include non-nested terms of a certain type (such as main
effects) for all combinations of a selected set of factors and covariates.
- Select one or more factors or covariates or a combination of factors and covariates.
- Select a method for building the terms and click Add.
- Repeat the process until you have all of the terms that you want in the model.

- Build custom terms
- Use this choice when you want to include nested terms or when you want to explicitly build any
term variable by variable. Building a nested term involves the following steps:
- Select a factor or covariate that is nested within another factor and click the move button.
- Click Within.
- Select the factor within which the previous factor or covariate is nested and click the move button.
- Click Add.
- Optionally, you can include interaction effects by using By, or you can add multiple levels of nesting to the nested term.

Specifying an Auxiliary Regression Model for GLM Univariate

This feature requires the Statistics Base option.

- From the menus choose:
- In the GLM Univariate dialog box, click Options.
- In the Univariate Options dialog box, select Modified Breusch-Pagan test, Breusch-Pagan test, or F test and click the associated Model button.