A report about this was published in the Journal of the AES in 2002 by Drs. Pastillé and Ochmann from Berlin, entitled "The 10 dB switch in condenser microphones."
Thanks to a taperjoe and a friend of his who is a member of the AES I was able to get a copy of this study and read it cover to cover last night. A good read, lots of math.
I think of the whole study the excerpt that strikes me as explaining exactly what we are trying to get at in this thread is here...
" If an audio engineer desires distortion-free recordings,
a careful consideration of the microphone circuit dia-
gram is necessary. A good option is to use a microphone
with a high level limit. The development of a micro-
phone amplifier with a maximum input voltage of 5.5 V
by a German manufacturer1 makes the discussion of the
10-dB switch obsolete. In this case a reduction of the
output signal is possible using a passive network.
However, it should be noted that the distortion can be uti-
lized for artistic purposes."
This plainly states that for distortion free recordings, DON'T use your mic pad, use the pad in the amplifier(or something inline), that's what they were made for. If the mic capsule is distorting, use the on-mic pad. Notice the main concern here isn't even noise floor(SNR). It's distortion, which IS a concern, much moreso than SNR according to the study. Thanks to DSatz for pointing me to the study, it was a good read, learned a ton.
Ok and, these guys look like they not only know what they're talking about, they look like some serious party animals, someone invite them to the next show!