It is simple,
The measurement are sensitive on frequency and it tooks quite expensive equipement (and it has to be regukarely calibrated) to make an acurate measurement.
Now a leatle bit of physics...
Look at the speaker an the microphone itself. If You neglect the electrical calculations at this moment and You just look on them of two bodies that have a mass. The force that moves the speaker is F = Ma, where M os a mass of the speaker membrane and a it's acceleration. If you apply simple sinewave oscillations of amplitude A on a membrane, You coud describe its position as a
X = A sin(omega t)
Here omega is a radial frequency, ant t is time. If You calculate a acceleration of the membrane (d^2 x / dt^2) You will get
a = -A omega^2 sin(omega t)
The power of Your amplifier is rated as some finite value. but let's stay at the definition of the power.
P = Work / t = F * x / t
that means in case of oscilla tions
P = ma * x / t = m A^2 omega^2 sin^2(t) / t
It is not a simple relation.
Imagine now, all of the power that You have applied thru the coil to the membrane is used to compress some volume of air near the membrane. and You have the reverse process on a mike.
You could see that in case that You have a finite power source the amplitude A of the oscillations of a membrane varies with the frequency. Just to have a leatle bit on complexity of a events that happend in such situation, imagina a mass of air as a spring with an friction element (sort of fluid brake that brakes with a force that is some factor multiplied with the velociti on power between 1 and two).
Everything is extremley sensitive on frequency. And all the story behind the boxes for the speakers and the division in two or more ranges (HI - LO, HI - MID - LOW, HI- MID - LOW - SUBSONIC) is a optimization for that complex system with a goal to achieve as flat frequency response as possible of a speaker system.
That is a reason that You could hear a tremendous difference between a bad and a good speaker (compare a sound of same aplifier on several speakers, I do not want to promote some manufacturers here but You will get an idea).
So it is all very complex system that we wanted to idealize and make as close as possible to a perfect one, it is, and it will always be a frequency (and a position in a ral room) dependent system. The rooms for the audio research and measurements need to be isolated from external noise and with damped reflected noise as much as possible.
Sincerely Yours
Nenad Sakan