This is, of course, totally incorrect. All scientific work relies on taking the visual bias out of an experiment. In my industry, we do this all the time. Of course, valid results assume appropriate

for statistical relevance. Our sampling panels are large, and I accept that this can't generally be done in an audio context. However, this one

 

is quite wrong, as that is eliminated by the test procedure. If one can't see what is being tested, there is no bias.

There's no way to put this gently, but, yes, they are. However, if they're all happy with their choices and believe they hear a difference, I for one am not going to object.