The actual effect of the placebo is usually at most a small portion of all the factors that are measured by it. Of course this depends heavily on the experiment in question and some drugs (like psychological ones) have a higher placebo effect than others.
A lot of people seem to have the idea that the placebo effect is very big for medication outside of experiments - but most of the time it's very small to non-existing.
I don't think that anything has been demonstrated regarding allergies, but I wouldn't rule out a therapeutic effect in this case. The absence of evidence is not an evidence of absence.
I went from not being able to eat anything with wheat or soy lecithin to having no problem with it:)
Two reasons: 1) bias is a sufficient explanation, 2) the prior probability for biases effecting experiments is huge, while the prior for beliefs having a strong effect on physiology (with the exception of highly subjective phenomena like pain and mood) is comparatively quite low.
>so surely in these trials the placebo effect is legitimate
It would superficially seem so, but the key point to remember is that cognitive biases are not something we consciously apply. They are instinctive heuristics that worked really well at helping us survive in the ancestral environment (long before the concept of empirical tests), and even when we know about them, we can't turn them off. Cognitive biases work a lot like optical illusions. You've probably seen the checker shadow illusion[1]. You can understand how the illusion works, and know full well that the two squares are the same color, and you can even watch an animation that proves it to you[2], but when you look at the final image, square A will always look darker than square B. Knowing about the illusion doesn't fix it.
So even placebo controlled studies do not allow us to be unbiased in our perception. What they do is allow us to measure the effects of our biases, so that we can compensate for those effects in our calculations.
[1] http://upload.wikimedia.org/wikipedia/commons/6/60/Grey_squa...
[2] http://upload.wikimedia.org/wikipedia/commons/9/93/Optical_i...
Using "measurement error" as an argument is kind of insulting to the testers as most of them have got to have a grasp of statistics and other factors in order to publish results that aren't going to get them laughed at by the scientific community.
How do you explain placebo effects in double blind tests?
It's not a yes or no answer. Part of what we call "placebo effect" is certainly the result of priming and confirmation bias, but for some specific subjective phenomena (like mood), it is plausible that the effect is real.
>Using "measurement error" as an argument is kind of insulting to the testers as most of them have got to have a grasp of statistics and other factors in order to publish results that aren't going to get them laughed at by the scientific community.
>How do you explain placebo effects in double blind tests?
Being aware of cognitive biases doesn't immunize you to them. There isn't even very good evidence that being aware of cognitive biases reduces your susceptibility to them. There is no shame in that, and any scientist who is insulted by the suggestion that s/he is susceptible to biases should go back to Science School. What being aware of biases does allow you to do is design experiments that compensate for those biases.
And that's the whole reason that we have double blind tests. We know that since humans have to administer and participate in experiments, a lot of biasing factors are going to come into play. So we have come up with designs like that which allow us to measure the extent to which the effects we appear to see are due to biases (or the power of suggestion), and how much the effects can be attributed to an actual effect of the drug.
>Over any period of time, a person will have a great many variations in mood, pain, etc.
>When [...] a subject or researcher has been primed to expect [an effect], the subject or researcher is more likely to notice that effect.
>to this extent, a placebo (or nocebo) effect can be accounted for [by these measurement errors.]
>It is plausible that some specific placebo effects like reduction/increase of pain response, anxiety, and mood could actually be real psychological effects[.]
This is just as likely to be true for the "nocebo" effect as for the placebo effect. Dizziness, for example, is a very common symptom for people to experience for myriad benign reasons. Orthostatic hypotension (head rush) is a common cause of dizziness, and it is affected by a lot of minor factors like hydration, vasodilation, how long you've been sedentary, how fast you stand up, etc. This may happen to you three times in a week, and you'd never pay it any mind. But if a doctor told you that a medicine you were taking might cause dizziness, you would notice it.
So to this extent, a placebo (or nocebo) effect can be accounted for with absolutely no "mind over matter" style physiological change whatsoever. It is plausible that some specific placebo effects like reduction/increase of pain response, anxiety, and mood could actually be real psychological effects involving suggestion. But if you hear of a placebo doing something and you think "how could a belief in a placebo possibly have that kind of effect", it's very likely that you are at least partly seeing mere measurement error.
[1] http://en.wikipedia.org/wiki/N-rays