Factual
What is the best way to handle NA values when performing a regression?
When will the quantiles graph for a regression model not look like a nice line of fit?
Can you compare the anova versus manova results? Aside from the multiple sections, is there really a difference in the calculations?
When, how, and why?
Why does the Residuals vs Leverage graph show such a blob of data?
Why do we use 4 as a rounding number in the robust regression?
At what point will you feel comfortable deciding that the dataset you are using for a regression has the right set of predictors in use?
Challenges
Are there better predictors available for obesity than those used in the chapter?
How can multilevel regression be used for either the obesity or mpg datasets?
Can you determine a different set of predictors for mpg that does not reduce it to simple government fiat?