> Or drinks where you cannot tell if they have any alcohol in them.
Proving this for a sample is pretty much impossible because you’re simply masking it and the sensory performance of humans is highly variable. You can get some imperfect approximation of blinding but it will be a compromise.
You can still control for this, by having the participants taste both drinks, in a blinded manner.
If they can still somehow taste the difference between them, then you change your mix so that they can't tell the difference or eliminate them from the study.
This may have other effects on the study, but at least you can be reasonably sure that no one can tell the difference between the substances being tested, except possibly by the effects they feel after consuming the substances in question.
> If they can still somehow taste the difference between them, then you change your mix so that they can't tell the difference or eliminate them from the study.
This has lots of problems, not least of which it is assuming you could afford to tweak your experimental protocol this way all in the recruitment phase.
Do you have any examples in the literature where this is done?
At least for the study in question they used the method of swabbing a negligible amount of alcohol on the rim of the glass to attempt to fool all participants into thinking they were getting alcohol.
The problem is people know how alcohol tastes. Have you ever had a non-alcoholic beer?