Another poster mentioned that sometimes blind tests are passed when people are forced to guess even when they do not think they can hear a difference.
do you mean passing by guessing(a logic but rare occurrence from statistical reasons? or do you mean someone thinking in casual listening that he can't tell a difference but will notice it in blind test?
I see no subconscious at play here in either case. math, or again the tool reason. different tools can do different things.
Someone mentioned that flawed short term memory was a likely culprit. Are there any other explanations? How about information on the thresholds of low volume hearing? Perhaps one experiment that could be done would be to determine whether there was agreement between what listeners perceive as silence under normal conditions, and what they perceived as silence when they were forced to choose between successive periods of silence, and very quiet sounds that were just on the edge of hearing. I guess this is similar to what you are talking about with training.
the serious tests conducted up until now tend to point at memory as the errors add up with time passing. if we listen again, sometimes we realize we had it wrong(so we correct our memory with fresh data, but often we will just confirm what we think we've heard before. that's why preconception and external biases are seen by us objectivists as being so dangerous when trying to find out the truth. once we get an idea, it's very hard not to focus on it to the point where we feel it for "real". ever had someone tell you there was a mosquito in the room? it doesn't matter if it's true, you will feel like you hear it and start scratching ^_^. the brain is amazing, but it comes with it's very own flaws.
about threshold and silence, the kind of things you talk about will change depending on what we're testing. like listening to quiet noise and listening to quiet noise while louder noise are played, that wouldn't give the same results. lossy music is built on that understanding, masking, how we don't have the same sensitivity when loud sounds are played it's not just a brain thing, we have acoustic reflex in the ear that physically reduces our sensitivity. so with what you're suggesting, you need to be very clear on what you're testing and what you can conclude from the results. for dynamic around 60db is what seems accepted as our instantaneous limit( SE come back you say it better than I do). for music playing, with all the quieter passages and situation when something might not be masked, you're usually good with about 80db. and you really need special conditions to get beyond that.
but again you focus on quiet stuff because you feel logically that if it eluded us in a blind test, it has to be a quiet part at play. but from a physiological standpoint, it would seem more rational to me that the small changes in the loudest sounds would be what can impact something such as fatigue. it's a matter of energy. like how we have a hard time noticing a 0.2db difference between 2 files, but it's enough to make us "feel" like one is better. my personal intuition would push me toward those stuff instead of thinking that the quiet sounds matter more than they do. but I have no evidence one way or another, so I'm not saying my idea is better than yours. just why I think that way.