ABSTRACT
Multiple regression is the oldest and most well established of the modelling techniques and is appropriate when the dependent variable is continuous. Explanatory (independent) variables can be continuous and/or categorical. Multiple regression is closely related to analysis of variance (ANOVA) and analysis of covariance (ANCOVA). In fact ANOVA and ANCOVA are alternative but complementary ways of presenting the same information and, carried out on the same data, will give exactly the same inferences as multiple regression. ANOVA is appropriate when all the independent variables are categorical and ANCOVA is appropriate when the independent variables are both categorical and continuous. As they are alternative ways of presenting the same results, most statistical computer packages, including SPSS and Stata, present analysis of variance results as part of the output from multiple regression. As multiple regression more readily conforms to the formulation of the general linear model (see Equation (1.1)) I feel that for most applications multiple regression is preferable. However, one or two statistics presented as part of the ANOVA results are helpful in interpreting the regression model.
