The success rates of two treatments (A and B) for clinical depression are being compared. The research team included five doctors, and the participants were 200 patients with depression. The doctors were supposed to randomly assign treatments to patients, but two doctors didn't do this. Instead, they often assigned more severely depressed patients to use treatment A and less severely depressed patients to use treatment B.

a. Three variables measured by the investigators were the outcome of treatment (successful or not), method of treatment (A or B), and a rating of the initial severity of the depression (mild or severe). Which of these variables is the response variable in this investigation?

b. The doctors were surprised by the final result, which was that treatment B had a higher success rate than treatment A. Based on results attained by other investigators, they had expected the opposite to occur. Explain how the actions of the two doctors who did not always randomly assign treatments may have caused an unfair comparison of the two treatments.

c. Taking given Example into consideration, describe how the researchers might control for the effect of the initial severity of patient depression when they analyze the relationship between treatment method and outcome.

Example
Blood Pressure and Oral Contraceptive Use This is a hypothetical example of an observational study done to examine the association between oral contraceptive use and blood pressure. Although the data are hypothetical, they are similar to data from several actual studies of the same problem.