> 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.

> 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.