Rawhide commented on my previous post that he was surprised there was much doubt about the placebo effect’s existence. It turns out there’ve been some serious questions raised about whether the placebo effect is actually just a myth after a 2001 New England Journal of Medicine article that analyzed 114 placebo-using medical studies and found that, on the average, the placebo effect was minimal if it exists at all.
Dylan Evans (a frequent contributor to the MindHacks blog) has a 2003 book called Placebo: The Belief Effect that argues that the placebo effect only helps with some kinds of conditions — namely pain, swelling, stomach ulcers, depression, and anxiety — and that by lumping everything into one average the meta-study washes out the few places where placebos actually work. He also suggests that placebos probably work by triggering the release of endorphins, which affect the same kinds of symptoms. Given the recent study it looks like he hit the nail on the head on that one.
You can find a nice summary of his idea in this short paper, which also includes a nice history of the discovery and our understanding of the placebo effect.
Update 8/30/05: typo fix