Lately I've been measuring at 4 meters just because Karlsons don't converge up close, and I don't imagine a reflex does either, even when doing just one driver. I've had the most luck with ground plane, especially indoors, and I've learned to leave the unfixable room anomalies alone, rather than try to take them out with parametric EQ. All of my graphs show a high-Q suckout at 62hz and a broader hump at 160hz even after throwing rolls of insulation, pillows, futon, and a mattress around. Heck with it, it's there.

The best place for measuring in my house is the living/dining area that opens to an upper level, so there are the usual constraints on how long I can leave a rig set up.

You mentioned reasonably assuming the level is unchanged when going to pink noise from sine - I agree, but it's kind of a leap of faith that I'm not willing to take without corroboration, and is really what started this. TruRTA has a calibration routine, but I don't like the way it works - the graphs don't match my SPL meter.