To illustrate a switching replication with treatment removal design, consider a study testing an exercise intervention on nonequivalent groups of depressed patients and depressed students. Researchers first measure baseline depression for both groups. Then, the patients begin the exercise program while the students do not. A second measurement should show decreased depression solely in the patient group. In the critical next phase, the treatment is removed from the patients (they stop exercising) and applied to the students (they start exercising). A final measurement should reveal that the students' depression levels have decreased, while the patients' depression levels have increased because they are no longer receiving the intervention.