Most codes of research ethics and the practice of Institutional Review Boards (IRBs) allow human subjects to withdraw from research at any time. Consent forms invariably make a statement to this effect. So understood, a subject's right to withdraw from research is inalienable; she cannot, through her consent, surrender this right. Recently critics have argued that in selected circumstances the right to withdraw from research is alienable; subjects have the moral authority, through their consent, to obligate themselves not to withdraw. Two kinds of cases have been cited to support this. In one case, there will be great benefits lost if subjects are permitted to withdraw before the completion of the protocol. In the other case, there will be harm to third parties if subjects withdraw from the experiment. In this paper, I defend the inalienability of the right to withdraw from research. I argue, first, that securing the desired benefits and avoiding the feared harms can be achieved without allowing waiver. Second, I show that permitting waiver in these cases does not guarantee that the ends sought will be achieved. And third, I articulate positive reasons for conceiving subjects' right to withdraw from research as inalienable.