I measure, sometimes out of curiosity and also when I do shoot-outs. Some records are cut "hotter" than others and the nominal level is higher, which may tilt the preference.
To add to the variables that @lewm mentioned, I find that every record has a playback level that seems "just right" for me--too loud and it sounds like the program is being "played at you" and too low a setting, and the room isn’t energized at the bottom end (assuming there are deeper fundamental notes with a double bass, piano, etc.).
Room size is important too. I’ve mentioned this elsewhere, but the ability of a system to "scale" in size and dynamics is an important part of replicating a concert experience. I’ve got a big dedicated room, but it simply won’t replicate the sound of a 2,800 seat hall.
I tend to listen at modest levels because I can "get it" without cranking it hard--due in part to extreme efficiency of the loudspeaker and the nature of SET amps.
I usually use a couple different meters- one, that could be calibrated but I didn’t go to the expense, another as part of a set of digital tools that are on an iPad. Both are pretty close at a "C" weighting and peaks are a little over 80db. Which means that I’m nominally listening at lower levels.
I think there are so many variables among each system, room, program material and listener preference that it’s hard to say what the norm is. We have friends that own a big club in town. I was hanging at the board several years ago, and even with ear protection in (I always wear Etymotics at live events), the sound seemed crazy loud- I asked the guy running the board (shouting)-- 104db. Brain death.
