"African AIDS Patients More Diligent in Taking Medicine Than in U.S." was reported in 2003. On average, "American patients tell their doctors that they are doing 20 percentage points better than they really are-that is, a patient who says he takes 90% of his pills will, when tested with unannounced home pill counts or electronic pill-bottle caps, turn out to be taking 70%."46
a. If we regressed claimed percentage on actual percentage of pills taken, would the number 20 correspond to intercept, slope, or correlation?
b. The article mentions that "a study of 29 Ugandan patients found that, on average, they estimated that they were taking 93% of pills and proved to be taking 91%." If two regression lines for pills claimed versus pills taken were drawn on the same graph, one for Uganda and one for the United States, would the line for Uganda be higher than or lower than the one for the United States?