In this video we continue with Example #7 from Mod7Examps.sas, in which we regress the city mpg on weight for 93 cars in the 1993 Consumer Reports Cars dataset. Here we consider a model that is quadratic in weight. Such a model fits significantly better than the simple linear regression model that we fitted previously. We also show how to produce a predicted value at a given value of the explanatory variables (which is also the point estimate of the mean at a given value of the explanatory variables) and we demonstrate the dangers of extrapolation beyond the range of our data.