Multiple Regression Button

Click the button to display the Multiple Linear Regression Startup Panel. The general purpose of multiple regression is to analyze the relationship between several independent or predictor variables and a dependent or criterion variable.

The Multiple Regression module will perform least-squares multiple linear regression and compute detailed residual statistics. Methods for forward and backwards stepwise selection of predictor variables are also provided. Both intercept and non-intercept models can be evaluated, and different methods can be chosen for computing the overall ANOVA table and R-square statistic (for non-intercept models). STATISTICA will compute all standard multiple regression results statistics, and perform extensive residual analyses. Predicted values can also be computed for user-defined values of the predictors.

See also the General Regression Models (GRM) module for complex linear models, best-subset selection, and other advanced techniques; see the General Linear Models (GLM) module to fit complex linear models; see Generalized Linear/Nonlinear Models (GLZ) and Nonlinear Estimation to fit nonlinear models; see also Generalized Additive Models (GAM), Partial Least Squares Models (PLS), and General Discriminant Analysis Models (GDA) for more specialized procedures.