Regression Episode 3: Testing the Model
Dan continues his discussion of the regression model with an exploration of methods for making inferences about the overall fit of the model. He begins with a brief review of the simple one-predictor case with an example of the regression of performance on experience...
He then walks through the computation of sums of squares, degrees of freedom, and mean squares, which are used to calculate the multiple R-square for the model (variance explained) and the overall model F-test (whether the variance explained significantly differs from zero). He concludes with a computer demonstration in which he examines the prediction of road sobriety test scores by number of beers consumed.