???Why The HUGE Discrepancy in Sensitivity Ratings???

FYI, sensitivity and efficiency are two different things, and is rarely the same for any given speaker.

Efficiency is measured at 1 Watt, and is also called "Power Efficiency." Sensitivity is measured at 2.82 Volts and is also called "Voltage Sensitivity." If the speaker is purely resistive and measures exactly 8 Ohms, then these two figures will be the same.

Efficiency explains how much work is required to produce a certain volume level, while voltage sensitivity cares nothing for the amount of work, and only thinks about what happens when a certain voltage is applied. Take for instance a 1 Ohm speaker, remember the old Apogees?

At 2.83 Volts, an 8 Ohm speaker draws 1 Watt, a 4 Ohm speaker 2 Watts, 2 Ohm speaker 4 Watts, and a 1 Ohm speaker draws 8 watts.

Sensitivity could be accurately called 85 dB, but at 1 Ohm it would actually be drawing 8 Watts, so the efficiency is actually a lot lower, probably around 78 dB or so.  I've gotten quite forgetful about the precise math, but you get the idea.

This matters more to tube amplifiers, which have the same power output regardless of the transformer taps. For those with reasonably beefy solid state amplifiers, and speakers with impedance curves at 3 Ohms or higher voltage sensitivity matters more.



Best,


Erik

freediver,
i'm looking at the M-L 35XT speaker manual & on the last page, the specs read:
Sensitivity = 92dB @ 2.83V / meter.
You can find the manual here:
http://www.martinlogan.com/pdf/manuals/manual_motion_35xt.pdf

The speaker is clearly noted as a 4 Ohms speaker.
If you input 2.83V into 4 Ohms (after you calculated V^2/R) you get an input of 2W.
Usually the standard is to input 1W into an 8 Ohm speaker. 
So, for the M-L Motion 35XT we have to compensate/normalize to bring Sensitivity number to an equivalent 8 Ohms so you can do an apples-apples comparison with other speakers.
We need to halve the number of input watts to 1W Or double the input speaker impedance to 8 Ohms (which will also yield an input of 1W).
In either case, the output SPL should drop by 3dB (assuming the speaker SPL response is linear in this region & is being fed by power well within its handling capacity. This should be the case as 1W is well within the power handling capacity of the speaker which also means its not distorting hence response should be linear).
So, in effect the M-L Motion 35XT Sensitivity is 92-3 = 89dB.
You say it's measuring 87dB ( I could not find a review that said this. Can you please provide a link? Thanks).
it's not out of the ordinary to lose 2dB due to measurement setup issues. And, like others wrote, the spec could be in-room & the measurement could be in an anechoic chamber. 
Anyway, you should view the M-L Motion 35XT as an 89dB sensitivity speaker; not 92dB. 

These last posts from timlub, erik_squires and bombaywalla should be carved in stone.

I have a website. I have a frequency response section. I come right out and tell everyone: if your stupid enough to believe these figures, I won't sell you a pair of speakers.
 
We've all been lied to for so long, we actually want to read it, and believe it.

A day will come, when we let our 'EARS' determine how good a speaker is, and not the shite a marketing department exhales. Don't get me started on 'reviewers', and there single minded endeavor (endeavour for us Canadians) to garner more advertising dollars. We're a long way away.

Listen to the speakers. If they don't put the performers in the room with you, cross them off your list. Marketing garbage.

I agree with nonoise.