Randomized trials depend on clinicians feeling that they are morally justified in allowing their patients to be randomized across treatment arms. Typically such justification rides on what has been called "clinical equipoise"--when there is disagreement of opinion among the community of experts about whether one treatment is better than another, then physicians can ethically enter their patients into a clinical trial, even if individual physicians are not at equipoise. Recent debates over prominent studies, however, illustrate that controversy can be easily created rather than dispelled by trials, with many clinicians choosing not to use the proven therapy until they receive more convincing evidence of its superiority. In such situations, we propose that a new standard of equipoise be used to guide decisions about the ethical justifications for research trials--a standard of behavioral equipoise. Under behavioral equipoise, a trial is potentially justifiable if it addresses behavioral resistance to prior scientific evidence.