This enables you to change the current regression model by adding or removing terms. The Options settings of the regression menu (or GLM menu) control what output is produced when the model is updated. After fitting a sequence of models you can use the Accumulated setting of Further output to produce a summary of the various changes you have made. You can also use this menu to build regression models using a stepwise search.
Specifies the terms (variates or factors, or interactions) to be changed. You can list either identifiers or a formula specifying a more complicated set of terms. If new terms introduced have missing values in units not already excluded from the analysis, then the sequence of models will be interrupted because the analyses will not be comparable. The previous model will be refitted with the new set of units, and then the new one will be fitted.
The specified terms are added to the current model. The current model is updated to contain these terms.
The specified terms are removed from the current model. They must be present in the model currently fitted. The current model is updated to reflect terms which have been dropped.
Terms that are present in the model are dropped, and terms that are not present are added. The current model is updated to reflect the changes.
The effects of changing the model is reported, for each of the terms in turn. Terms that are present in the model are dropped, and terms that are not present are added. However, the current model is retained as before.
You can produce a brief summary of potential changes by selecting Display changes. You can control whether the previously chosen display options are printed in the analysis by setting the Display default options checkbox.
Max number of steps
Specify the limit on the number of times to repeat a stepwise selection, unless no change in the model is made.
Specify the criterion for the inclusion and/or exclusion of terms. The criterion is based on the variance ratios. As the criteria are ratios of variances, they should not be interpreted as F-statistics with the usual interpretation of significance.
Builds a regression model using forward selection. You can specify the terms to use in the selection process in the Terms field.
Builds a regression model using backward elimination. You can specify the terms to use in the elimination process in the Terms field.
Builds a regression model using stepwise regression. You can specify the terms to use in the selection process in the Terms field.
If selected, all changes made to the model will be displayed in the output.
If selected the progress of the fit of the model will be displayed in the output.
Display default options
If selected, all display options previously selected will be displayed in the revised analysis.