In my job I have qualified examination procedures to DBT where we lead a team of technicians through task of identifying flaws in components according to strict procedure instruction. First the procedure is tested and secondly the personnel are tested, thus double blind in my application. This is only partially related to the discussion here but I have a couple of observations after leading many of these exercises.

To have more than one person select the same attribute with accuracy, both need to be trained similarly, a point already made in this forum. For audio, this might require the test takers to be, at the minimum, passionate listeners as opposed to dispassionate, which most Americans are unfortunately ,Test participants need to be round the same age. Hearing loss between say 55 and 70 could be significant and skew results.

I thought improvements in equipment came from careful selection of components and attention to wiring and assembly not dbt. Anyway, I'm one of those guy that hears minor improvements from cable, DAC upsampling and vibration control. I must be cursed.