GDA Models Results - Regression Residuals 2 Tab

Select the Regression residuals 2 tab of the GDA Models Results dialog to access options to display residual statistics and predicted values. The options on this tab allow you to flexibly combine into histograms, scatterplots, etc. various regression-like residual statistics. As described in the Introductory Overview, the GDA module applies the general linear model to the discriminant function analysis problem, after (internally) recoding the class (group) memberships recorded in the categorical dependent variable into multiple (dummy-) coded dependent variables.

Dependent variables

Click the Dependent variables button to display the Classes for Residuals dialog, in which you select the classes of the dependent variable to be used for the residual statistics. The predicted and residual values are computed for the dependent variable's classes (or levels).

Sample. Select the respective option button in the Sample group box to specify which type of sample to base the predicted and residual values. You can display and plot predicted and residual values for all observations that were used to compute the current results (select Analysis), all observations that were not used to compute the current results, but have valid data for all predictor and dependent variables (select Cross-validation (Cross-val.)), both the analysis and the cross-validation samples (select Both), or display and plot predicted values and related statistics for all cases that have valid data for the predictor variables, but missing data for at least one dependent variable (select Prediction). If these options are dimmed, no valid cases in the prediction sample were found, and/or no cross-validation sample was specified on the GDA Quick Specs Dialog - Advanced tab, or via the SAMPLE keyword in the GDA Models Syntax Editor dialog.

X (var/pred/res)

For plots requiring a single variable (e.g., histograms), select the respective type of predicted value, residual value, or variable in the X (var/pred/res) list.

Y (var/pred/res)

For plots requiring two variables (e.g., scatterplots), select an additional type of predicted value, residual value, or variable in the Y (var/pred/res) list.

Z (var/pred/res)

For plots requiring three variables (e.g., surface plots), also make a selection from the Z (var/pred/res) list.

Histogram of selected X (variable, predicted, or residual value)

Click the Histogram of selected X button to display a histogram for the variable, type of predicted values, or residual values, as selected in the X (var/pred/res) list.

Probab. plot of selected X.

Click the Probab. plot of selected X button to display a normal probability plot for the variable, type of predicted values, or residual values, as selected in the X (var/pred/res) list.

Half-normal

Click the Half-normal button to produce a half-normal probability plot for the variable, type of predicted values, or residual values, as selected in the X (var/pred/res) list. If the Plot absolute values check box is selected on the Regression Residuals 1 tab (next to the Half-normal option) then the respective absolute values will be plotted.

Detrended

Click the Detrended button to produce a detrended normal probability plot for the variable, type of predicted values, or residual values, as selected in the X (var/pred/res) list.

Scatterplot of selected X vs. Y (var., pred., or res. value).

, pred., or res. value). Click the Scatterplot of selected X vs. Y button to produce a 2D scatterplot for the variables, types of predicted values, or residual value as selected in the X (var/pred/res) and Y (var/pred/res) lists.

Surface plot of selected X, Y, & Z (var., pred., or res. value).

, or res. value). Click the Surface plot of selected X, Y, & Z button to produce a 3D Surface Plot for the variables, type of predicted values, or residual values, as selected in the X (var/pred/res), Y (var/pred/res), and Z (var/pred/res) lists. The plot will show the raw data points for the plot, and the fitted quadratic surface, if a quadratic surface can be fitted to the data.