Also, and sorry if I'm misread what you are trying to say, but are you arguing that objectivists (even hard objectivists) are dismissing results of controlled listening tests?
I'm not sure, it is difficult to know exactly. There are people that say that a persons ears can deceive them; if you have 5 people who hear a similar thing, they have the same answer - likely they are being deceived; 50 people, the same answer.
If you stick those 50 people in a controlled, double blind test and the outcome is the same, 50 (or close to) heard what they claimed to have heard outside the test, suddenly this is considered valid and useful information?
I dunno, it seems to me the person who says that ears are easy to fool outside of a controlled background, might not have to high an opinion of using listeners within controlled test environments, although they'd rather not admit it.
Of course, it has to be controlled if the data is to have value from a scientific perspective, but it would be interesting to see the relationship between what people think they hear when in an uncontrolled setting and what they think they hear in a controlled one. I think the relationship between the two may be tighter than some would imagine. Just speculation on my part.
Theoretically in terms of logic this is actually not correct. It's still merely correlation, not causal, in that it can still be imagination but just imagination which happens to confirm the objective data. I also have this impression my current DAC of which I know measures better than the previous one, produces 'better' sound. Until I do a blind listening test of both, I cannot be sure there's really something to hear there, or just imagination.
Since you mention correlation, this one is always fun:
https://tylervigen.com/spurious-correlations
Right, I suppose in terms of logic, you are correct. However if a number of tests are done over a period of time and similar results are achieved, it becomes less and less likely that is pure luck/imagination on the part of the person taking the tests. Like you said, doesn't make it causal, but the likelihood of chance alone goes down.
There are quite likely people that have 'golden ears', not just in their own opinion, but in that they can discern bad reproduction from good just through listening. Likely not as well as a machine, but better than a high percentage of others.
I think these people probably could be identified by some kind of test, given a number of times over an extended period.