In a class experiment at Williams College, students were randomly assigned either to drink or abstain from drinking an energy drink before attempting a straightforward but tedious mathematical task. They were also randomly assigned to start the task at 11 P.M. at night or at 8 A.M. in the morning. (The resulting data set is available on the book’s website.) When we regress y = time it took to complete task (in minutes) on e = energy drink (1 = yes, 0 = no) and d = daytime (1 = 11P.M., 0 = 8A.M.) we get the prediction equation

ŷ = 9.8 - 0.39e - 2.0d.
a. Interpret the effect of the energy drink.
b. Interpret the coefficient of d.
c. State a corresponding model for the population and indicate which parameters would need to equal zero for the response variable to be independent of whether one drinks an energy drink, for given time of the day.
d. Fit the regression model with software and comment on whether the mean times differ significantly between consuming and not consuming an energy drink, for a given time of day.