SAVE Subcommand (GLM: Univariate command)
Use SAVE
to
add one or more residual or fit values to the active dataset.
- Specify one or more temporary variables, each variable followed by an optional new name in parentheses. For a multivariate model, you can optionally specify a new name for the temporary variable related to each dependent variable.
-
WPRED
andWRESID
can be saved only ifREGWGT
has been specified. - Specifying a temporary variable on this subcommand results in a variable being added to the active data file for each dependent variable.
- You can specify variable names for the temporary
variables. These names must be unique, valid variable names. For a
multivariate model, there should be as many variable names specified
as there are dependent variables, and names should be listed in the
order of the dependent variables as specified on the
GLM
command. If you do not specify enough variable names, default variable names are used for any remaining variables. - If new names are not specified,
GLM
generates a rootname by using a shortened form of the temporary variable name with a suffix. For a multivariate model, the suffix _n is added to the temporary variable name, where n is the ordinal number of the dependent variable as specified on theGLM
command. - If more than one
SAVE
subcommand is specified, only the last subcommand is in effect.
PRED. Unstandardized predicted values.
WPRED. Weighted unstandardized
predicted values. This setting is available only if REGWGT
has been specified.
RESID. Unstandardized residuals.
WRESID. Weighted unstandardized
residuals. This setting is available only if REGWGT
has been specified.
DRESID. Deleted residuals.
ZRESID. Standardized residuals.
SRESID. Studentized residuals.
SEPRED. Standard errors of predicted value.
COOK. Cook’s distances.
LEVER. Uncentered leverage values.