If a randomized controlled trial includes just two groups, one which gets a treatment and another which gets placebo, it cannot attribute any results to the placebo, because it's missing a control for the placebo, i.e. a group which gets no intervention at all. Many studies have made this error, but that doesn't mean there's never a placebo effect.

http://www.dcscience.net/2015/12/11/placebo-effects-are-weak...

the placebo effect is both larger and smaller than people intuitively assume.

it's larger in the sense that part of it is essentially impossible to control for; people in a clinical trial are predisposed to, at a minimum, think about the issue which is being investigated, which likely causes some sort of change in their mindset and subsequently their body. there's no real way to measure that adequately.

but it's smaller in the sense that the body isn't a system like in the Matrix where "the mind makes it real". if someone tells you that you drank lethal poison (but in actuality you drank a placebo), you might feel bad, but you won't die. the reverse is not necessarily true, bizarrely.

if you get a (unbeknownst to you) placebo/sham knee surgery as part of a trial to investigate the merit of a genuine surgical procedure designed explicitly to reduce pain, it'll reduce your pain almost as effectively as "the real deal". but here's the clincher. if you are told beforehand that it is a placebo, it'll still reduce your pain, although if i am remembering the study correctly it was a smaller benefit than those who were not told they received a placebo.

so if you know you're getting a placebo, the effect is WAY larger than it "should" be. but that's for a subjective variable -- pain. but when the subject isn't prompted about the effects of the placebo, or is administered a totally implausible placebo which can't possibly fix the issue -- like a pill to cure a severed arm -- it's even more impotent than we would expect.

I am not sure why you have been downvoted. Yes the placebo effect is often just a regression to the mean of a condition that varies. Imagine that you have a disease that goes up or down in serverity (say depression). People enter the trial when most ill and over the course of the trial they trend back to symptomatic mean (i.e. they get better on their own). This is why doctors favourite treatment "let's wait and see" is so effective.