Consider a well-designed randomized clinical trial where a cholesterol-reducing drug is being tested on humans. Cholesterol is measured as the ratio of the subject’s cholesterol level relative to a healthy cholesterol level—in percentage terms; so, a rate of 100% equates to a healthy cholesterol level. The treatment and control groups begin with the same average rate of 200%; thus, their cholesterol levels are 2 times the healthy level.
For a month, the treatment group receives the new drug and the control group receives a placebo; then at the end of the month, the cholesterol levels are measured again and found to be 175% for the treatment group and 190% for the control group. (8 points)
(i) What is the difference-in-difference estimate of the effectiveness of the drug at reducing cholesterol?
(ii) What is the benefit of a difference-in-difference estimate?