GDA Quick Specs - Quick Tab
Select the Quick tab of the GDA Quick specs dialog box to access the options described here. The specific options that are available on this tab depend on the Type of analysis selected on the General Discriminant Analysis (GDA) Models Startup Panel - Quick tab.
- Variables
- Click the Variables button to display a standard variable selection dialog box. Depending on the Type of analysis selected on the General Discriminant Analysis (GDA) Models Startup Panel - Quick tab, STATISTICA will prompt you to select a categorical dependent variable, one or more categorical predictor variables (grouping variables or factors in the design), and/or one or more continuous predictor variables (covariates). For example, if you selected Traditional discriminant analysis from the General Discriminant Analysis (GDA) Models Startup Panel - Quick tab, you will be prompted to enter one categorical dependent variable and one or more continuous predictor variables; if you selected General discriminant analysis from the General Discriminant Analysis (GDA) Models Startup Panel - Quick tab, you will be prompted to specify a dependent variable, and both categorical and/or continuous predictor variable lists. For details concerning different types of analyses, and the distinction between continuous and categorical predictor variables, see the Introductory Overview ; see also the Note of caution for models with categorical predictors, and other advanced techniques to learn about the possible limitations of including categorical predictors in discriminant analysis problems.
- Dep. var. codes.
- codes. Click the Dep. var. codes button to display the Select Codes for Dependent Variable dialog box, which is used to select the codes identifying the levels for the dependent variable. Codes must be integer values or text labels (but can be dates, times, etc.), and at least two codes must be specified for the categorical dependent variable.
- Factor codes
- Click the Factor codes button to display the Select Codes for Indep. Vars (Factors) dialog box, which is used to select the codes identifying the levels for the categorical predictor variables (grouping variables). Codes must be integer values or text labels (but can be dates, times, etc.), and at least two codes must be specified for each categorical predictor variable. This button is only available if categorical predictor variables (grouping variables) were specified for the current analysis via the Variables button (see above).
- Effects in design
- Click the Effects in design button to display the GDA Effects in Design dialog box. This dialog box is used to specify a custom design for the categorical and continuous predictor variables. This button is available only if General discriminant analysis is the Type of analysis selected on the General Discriminant Analysis (GDA) Models Startup Panel - Quick tab. If you have selected Traditional discriminant analysis as the Type of analysis and decide that the design needs to be customized, you can either choose the Syntax editor button on the GDA Quick specs dialog box to further customize the model via GDA Syntax, or you can return to the GDA Startup Panel and select General discriminant analysis.
- Error message: A column in the design matrix has no variance
- On occasion, the analyses in GLM, GDA, GSR, etc. (all modules where design matrices are analyzed) may not run to completion, because of the error A column in the design matrix has no variance. The design matrix is computed from the actual design points, using either sigma-restricted or overparameterized coding. If the design is mis-specified, given the available actual design points, then it can happen that, for example, for a coded interaction effect, all values are equal to a constant.
For example, consider this data file:
Suppose you specify a design that includes the Gender-by-Income interaction; the vector in the design matrix that represents that interaction will contain the same constant for all cases because the interaction effect in this incomplete design is not defined.
When it is not clear why this error message occurs, compute crosstabulation tables using Basic Statistics to determine the structure and distribution of design points over all possible combinations of factor levels, and specify the design accordingly (i.e., omit factor effects that are clearly not defined).
For alternative ways of specifying designs in GDA, see Methods for Specifying Designs. See also the GDA Index.