Alex, in his blog post from earlier today, makes a good point about placebos. Sometimes the patient is getting better anyway, and we should not attribute this effect to a placebo.
Note, however, that the best-known “anti-placebo” study is not as strong as is commonly believed. It relies heavily on a meta-analysis of other studies. Placebos appear to be effective in relieving the sufferer of pain, if nothing else. And placebos appear more effective when the ailment is continuous rather than discrete. Furthermore it is unclear how many people in the so-called no-treatment groups in fact received no treatment at all.
Robert Ehrlich’s Eight Preposterous Propositions offers a very good survey of the placebo debates. His conclusion:
In summary, the [critical] study may have shown that the placebo is not as powerful as some observers would believe, but it certainly is far from powerless.
By the way, did you know that people can become addicted to placebos, or suffer from harmful “side effects”? I’ll try to write more on “nocebos,” or negative placebos, soon, at least provided that my mental attitude holds up.