Among the BRFSS participants, the linear relationship between body-mass index (BMI) BMI and drinking frequency is to be compared between those who reported any exercise in the previous month (Exercise 5 1) and those who report no exercise (Exercise 5 0). Output is provided for separate straight-line equations among exercisers and non-exercisers, as well as for single models that include exercise, represented as a dummy variable.

a. Using method I (comparison of two lines), do you believe the lines are coincident? (Hint: Only one hypothesis test is necessary.)

b. Using method II (single model), do you believe the lines are coincident?

c. Do the tests of coincidence agree? Based on the magnitude of the results, is this expected? d. Using either method, test whether the lines are parallel.

e. Using either method, test whether the lines have the same intercept.

f. Comment on the validity of the homogeneous variance assumption for these data.