Answer the following questions about the rabbit experiment in the previous problem.
a) It is not possible to use an interaction term in this situation. Why not?
b) The additive linear model can be fit with the predictors given in either order. Compare the ANOVA tables that result from doing it in each order. How and why do they differ?
c) Explain how to obtain two of the four p-values (two in each ANOVA table) using a model comparison test.
d) Which analysis is the correct analysis for this situation? Most importantly, which p-value should we use to check for a possible treatment effect? Why?