The effect of a training program concerned with educating heart attack patients to the benefits of moderate exercise was examined. A group of recent heart attack patients was randomly divided into two groups; one group received the training program and the other did not. The dependent variable was the amount of time taken to jog three laps, with the weight of the patient after the program used as a covariate. Examination of the data after the study revealed that the covariate means of the two groups differed. Which of the following assumptions is most clearly violated?
a. Linearity
b. Homogeneity of slopes
c. Independence of the treatment and the covariate
d. Normality