That type of switch makes the microphone itself less sensitive--it reduces the microphone's output levels by 10 dB. But every microphone has a certain amount of residual noise and the switch has no effect on that. So if you use the switch when you don't absolutely have to, you're making your recording 10 dB noisier than necessary.
The real question is, when must you use the switch? The only time you have no alternative is when the electronics of the microphone's are in danger of overload due to very high sound pressure levels (or sometimes by wind). Microphone capsules can withstand deafeningly high sound pressure levels as a rule, but the circuitry can't, so the "maximum SPL" specification for a microphone describes what its circuitry can accept. If the specs say that that's less than 120 dB SPL, then yes--if you close-mike certain types of sound sources, occasionally something will get above that level. If the mike can handle (say) 125 - 130 dB SPL or even higher, then that risk is lessened considerably.
But 120 dB SPLs never happen "out in the room" where I think most people here do most or all of their live recording, unless you're in a club that is pushing the P.A. system to levels that are extremely dangerous to people's hearing AND you're fairly close to the speaker(s)--in which case I hope to hell that you're wearing very good hearing protection at all times while in that environment.
With most consumer and "semi-pro" recording equipment, even at relatively loud shows it's more likely that the mike inputs of the preamp or recorder will distort than that the microphone itself will be overloaded by high sound pressure levels. So a resistive pad (placed at the input of the preamp or recorder) is more often required than the "pad" switch on the microphone itself--and the nice thing about those resistive pads is that, unlike the pad switch on the microphone, they don't make your recordings noisier.
By the way, you're not alone: No human being can estimate high sound pressure levels accurately. The way our hearing works, the sensation of loudness is highly frequency-dependent, highly dependent on the duration of the stimulus, and there's also a kind of "built-in subjective compressor" that comes into effect--and you can't really tell when it is coming into effect! No one who knows elementary psychoacoustics would try to gauge sound levels above, say, 90 dB SPL by ear. To do so below that level takes practice, and the person would still be wrong a lot.
--best regards