First off, good job on looking at something unusual! When I first read the news report that led me to your paper, I really smiled at the novelty of performing a study on placebos. The reason for my comment is that I have a suggestion about some of the limitations you describe. You state it’s impossible to control for report bias, but it seems to me that there are a few simple things you could do to control for those effects. Adding a third group to this study (which others have suggested in the comment section as well, though just to see the comparison) of hidden-label placebo recipients might be a good place to start. The key would be to phrase the prescription of the placebo in a way that would be likely to encourage the participants of the third group to “please the experimenter” to an equal degree as those who receive the open-label placebo might. You could state that “This is the first trial for efficacy of this drug since being declared safe for human use,” or something to that effect.

As for those who may be disappointed with their assignment, I’d think downplaying the effects of the placebo, even deriding them, might make them feel better about not being selected. Basically, how you sell the treatment or lack of treatment to each group can be used to control their excitement or disappointment accordingly. I also think that you should sterilize the language of the next ad you place as well. Calling it a “mind-body” intervention seems awfully loaded and likely to draw in those seeking new age spiritual treatments. Maybe you should just sell it as a regular drug trial?

Well, those are just my two (or three) cents. Whatever you feel about the others, the suggestion I’d most like to see implemented in your next study would be the inclusion of the 3rd group. Best of luck to you all and thanks for the original thinking!