some people can't hear 16Khz very well, you need to compare it to that same person's graphs of different phones.
this is a great reference for yourself, but between people.. not so much. and of course, those with less background noise will have a different volume, which also alter how well we can hear. infact thats why everyone's bass response is bad, because music is designed to be played at >75db, and everyone here is most likely testing at like 65db unless theres lots of things going on around them, but they are probably testing alone quietly to try to get something accurate. i'd say for reference, listen to your favorite music at your normally loud level, and then choose 1khz that sounds about that loud.
i think a unified standard for testing FR is most important between people. while to one person, testing it like this is worthwhile. unless i am missing something (we pick a reference sound, and try to make all the other tones sound the same loudness?)
also, keep in mind that recordings typically have louder treble as they are designed to be played from loudspeakers, where high frequencies get absorbed in the air.
and personally, i find it hard to judge the loudness of 16khz in comparison with the rest - and i dont think there would be any correlation between us in that respect. one way to get rid of that is to do it as, well, a hearing test. that would significantly objectify it especially for ourselves. then we would have to apply the loudness U curve to it.
and yeah, just to throw it out there, not only does hearing vary differently between everyone at different frequencies, but how they are perceived when the volume goes up and down the same amount varies a lot as well. if you double the volume for someone at 1khz, it may only be 1.5x for someone else. while for the same two people, at 10khz, the 1.5x and 2x could be reversed.
to reiterate.. this is an ok method when comparing headphones yourself and then posting them here, but should never be used to compare between people.