ABSTRACT
Although the bivariate regression model is useful for illustrating the technique of regression, it has significant limitations in practical applications. As with the bivariate regression model, the multivariate model requires that independent variables be interval level or dichotomous (i.e., restricted to only two values). The interpretation of the partial slope coefficient for a variable is the same whether the variable is dichotomous or interval level. The assumptions of the multivariate regression model are natural extensions of the assumptions of the bivariate model. Social scientists also may develop multiple regression models by identifying a dependent variable and attempting to develop a theory that can explain changes in the variable as fully as possible. This leads researchers not to focus on a small number of independent variables and treat the rest as statistical controls, but instead to construct a larger set of hypotheses about variables that influence their dependent variable.
