One would need to do repeated measurements with multiple same-model drivers, simultaneously, to control for QC / tolerance. Pretty straightforward.
No such thing as a more or less “disciplined” audiophile in the way OP mentions, that I can think of, since everyone is a sample size of 1 who’s more or less prone to acclimatization.
Amazing what a bit of background knowledge it science and stats can save you from having to just wonder. 😉