> I'm confused about why you feel that a large amount of the placebo effect is due to bias Two reasons: 1) bias is a sufficient explanation, 2) the prior proba

> I'm confused about why you feel that a large amount of the placebo effect is due to bias 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_square_optical_illusion.PNG http://upload.wikimedia.org/wikipedia/commons/6/60/Grey_squa... [2] http://upload.wikimedia.org/wikipedia/commons/9/93/Optical_illusion_greysquares.gif http://upload.wikimedia.org/wikipedia/commons/9/93/Optical_i...