Randomization is an important methodological strength of the RCT: It helps ensure the comparability of patients in the experimental and control conditions; it enables both patient and physician-researcher to remain ignorant as to assignment of conditions; and it guarantees the validity of statistical tests of significance (Friedman et al. 1998). Despite the advantages, randomization is the source of one of the most difficult and persistent challenges in the ethics of research. Until recently, most patients were offered enrollment in an RCT by their own physician; in many cases, the physician-researcher conducting the RCT was the patient’s physician. Physicians are widely regarded as having special ethical obligations to their patients. These include the obligation to ensure that the patient receives competent treatment. The problem is how can the physician, consistent with this ethical obligation, offer the patient enrollment in an RCT? As we have seen, in an RCT the patient will be randomly allocated to the experimental or control condition. On the face of it at least, tossing a coin to determine which treatment a patient will receive seems inconsistent with the physician’s obligation. Can the physician ever ethically offer RCT enrollment to his or her patient? If so, under what circumstances?