Because of the recent and controversial example of sham surgery for the evaluation of fetal tissue transplants for Parkinson's disease, there is renewed interest in the ethics of using "active" placebos in surgical trials, where otherwise there are no inert procedures available, and in pharmacological trials, where there are inert substances, but where patients may guess to which arm they have been allocated. This paper seeks to clarify the ethical arguments surrounding the use of active placebos in trials, and to set up a notation for assessing the ethics of trials more generally. We first establish an framework by which ethics committees can analyze such trials. We examine (1) the scientific value of the research; (2) the expected risks and benefits to individual patients, and (3) the voluntary nature of consent. We then contrast the implications of this framework for inert and active placebo-controlled trials, respectively. In particular, we analyze their relative expected utility using three main utility factors, namely, treatment effects, placebo effects, and altruism. We conclude that, when the intervention is already widely available, active placebo trials rely more heavily on altruism than do inert placebo trials and, when the intervention is restricted, this excess reliance may not be needed. What our analysis provides is the explicit justification for the apparent caution of Institutional Review Boards or ethics committees when reviewing sham operations, especially when the expected harm is not trivial and the risk of exploitation is high.