Why shouldn't there be a placebo effect? It's well documented for basically any drug, even for alcohol. Serve people non-alcoholic drinks but tell them otherwise, and they get drunk.
Because the trial wasn't about finding the expected effect (drunkenness) but the secondary effects. People don't get heart disease or cirrhosis from non-alcoholic drinks, or at least, I'd be interested in reading the study where they do.
No, but they can get short-term secondary effects such as nausea, headaches the next day, etc. Probably this category of things is what they want to control for.
