12 If the analysis-of-variance F -test leads to the conclusion that at least one of the model parameters is nonzero, can you conclude that the model is the best predictor for the dependent variable y ? Can you conclude that all of the terms in the model are important in predicting y ? What is the appropriate conclusion?