Question: a. Consider the data set analyzed in Section 15.1. Are there any other approaches (other than robust regression) that were presented in the preceding chapters that might be applied in analyzing the data? If so, analyze the data, using whatever method(s) seem appropriate, and compare your results with those given.
b. Is the normality assumption crucial considering Haiths objectives? Would having an approximately normal error term be necessary if Haith wanted to obtain prediction intervals?
c. Consider the additional concerns (beyond nonnormality) of Shapiro and Kiichner that were mentioned in the next-to-last paragraph of Section 15.1. What would you do before using any of the models that were discussed? In general, what action(s) should be taken regarding data that have been used in constructing a model before the model is used?