ABSTRACT
Behavioral research workers often conceive of regression in the limited sense of the linear, additive model in which the independent variables have a straight-line relation to the dependent variable and do not interact. The variables are categorial variables and continuous variables. In stepwise regression, a series of regression models is tried, each model including a different set of variables. The Regression models are simple regression, multiple regression, and stepwise multiple regression. Many researchers are more familiar with the analysis of variance (ANOVA) approach, and because in many situations regression analysis offers some advantages over ANOVA. Generalized regression has at least three advantages over ANOVA: the use of continuous variables, less data processing time, and direct, comprehensive estimates of the magnitude and significance of the independent variable effects. Regression of ungrouped scores not only analyzes all the accountable variance, but saves the investigator the bothersome task of establishing suitable cutting points.
