Blind testing works for components that have a wider spread of differentiation. It works for "preferences" as in which spaghetti sauce do you like. You can tell that immediately.
The trouble w extended listening is audio memory is the shortest memory we have and differences are more nuanced. There, "preferential comparison" (flavor of presentation over weeks or months) might be harder to assess except to say "I never noticed that before" in terms of sound articulation.
Even this cannot be written in stone as everyone’s brain is different.
What we hear is often compelled by desire of what we want to hear (expectation bias).
There is an aspect that adds further confusion to the process...chance. What would be an interesting experiment is if the aspect of "gamble" or "skin in the game" or "taking a chance" was eliminated because spending larger amounts of money on higher end stuff creates conflict of interest in my mind.
I cannot tell you the amount of times people coming into our salons, agonized over a decision when A/B (or C and D etc,) comparison differences were not as stand out between what they were listening to. Even taking things home for try out did not completely eliminate this conflict between money spent and objective/subjective auditory acquisition as the listener could not escape that big money was about to be spent at some point.
For the well healed, this aspect did not loom as large.