Options for MBUILD

Click the MBUILD button on the GLZ Analysis Syntax Editor - Specifications tab to display the Options for MBUILD dialog box, which contains options for the MBUILD keyword (see GLZ Syntax for a list of all keywords).

Element Name Description
Specifications Click a button in the Specifications group box to insert the respective specifications (option) at the current cursor location in the Analysis syntax edit field (in the GLZ Analysis Syntax Editor). The available specifications are described in detail below.
All The all option is used in conjunction with the MBUILD keyword; when all is specified, all effects in the DESIGN statement are automatically included in the model.
Forward The forward option is used in conjunction with the MBUILD keyword; the forward option will cause STATISTICA to perform a forward stepwise regression analysis, where effects (see DESIGN), once entered into the model, cannot be removed in subsequent steps.
Backward The backward option is used in conjunction with the MBUILD keyword; the backward option will cause STATISTICA to perform a backward stepwise regression analysis, where effects (see DESIGN), once removed from the model (see F2REMOVE, P2REMOVE, STEPCRIT) cannot be entered into the model in subsequent steps.
Fstepwise The fstepwise option is used in conjunction with the MBUILD keyword; the fstepwise option will cause STATISTICA to perform a forward stepwise regression analysis, where effects (see DESIGN), once entered into the model (see F1ENTER, P1ENTER, STEPCRIT), can be removed in subsequent steps if their contribution to the prediction of the dependent variable is not statistically significant, as defined by the F2REMOVE or the P2REMOVE keywords (see also the STEPCRIT keyword).
Bstepwise The bstepwise option is used in conjunction with the MBUILD keyword; the bstepwise option will cause STATISTICA to perform a backward stepwise regression analysis, where effects (see DESIGN), once removed from the model, can be reentered into the model in subsequent steps if their contribution to the prediction of the dependent variable is statistically significant, as defined by the F1ENTER or the P1ENTER keywords (see also the STEPCRIT keyword).
Bestsubset The bestsubset option is used in conjunction with the MBUILD keyword; the bestsubset option will cause STATISTICA to perform a comparisons of all possible subsets of effects (see DESIGN), and to find the best subset according to the specified criterion (see BESTCRIT) and given the respective subset sizes (see START and STOP).

For more information on model building methods, see Model Building in GRM in the Introductory Overview section.