GRAPHICAL INSIGHT INTO MULTIPLE-REGRESSION CONCEPTS

This article describes how pedagogical benefits can be achieved in the comprehension of (frequently) difficult multiple regression concepts by following Martin Gardner's injunction to simplify the problem. Specifically presenting the simplest multiple regression model (two independent variables) with the smallest nondegenerate data set (four points) and systematically changing the arrangement of the points, graphically one gains intuition into many important multiple regression concepts. These include the following: the size and determination of s2 = sigma2 , the size and meaning of R2, the significance of the overall regression model, the size and significance of each of the model's coefficients, multicolinearity, and how the location of the underlying data set influences each of these.