On the Zoom H4N, the “level” control goes from 0 to 100 (or is it 1 to 100?). What does it do actually? Specifically, is it controlling the degree of amplification of the incoming signal, or attenuating the incoming signal, or either around some true neutral setting?
There usually is a "native" setting called Unity. At that position, the signal is neither attenuated nor amplified artificially.
Basically you need to see what is the Unity for your particular device, and try to keep it at that level.
With 24bit recordings, softer sounds can always be successfully amplified/normalized in post.
Of course, you don't want to overload the input either - use built-in limiters if they are effective, or control the input via the external mixer/preamp.
In the absence of a separate "gain" control then the level control is probably attenuating the already amplified signal (considering the mics on that device are built-in so the amp input sensitivity wouldn't need adjustment for general use)/
The 0-100 represents the inverse of the signal attenuation.
I would guess that it amplifies the signal as this is the case most of the time.
An easy way to test this in any device is to record a tone (or anything with a standard comparable level) twice, once with a level of 10 and once with level 100. Take care to not clip the louder recording and to disable any effects, especially compressor / limiter. Then listen to both files but turn the volume down on the louder recording until the tone is heard at the same level. That way you will be able to listen to the difference in hiss/self noise of the preamp. If it is exactly the same, then the preamp has a constant value and the level is an attenuator. If the recording done on level 100 has more noise, then it is an amplifier.
In general, most portable recorders and preamps of that price range, are rather noisy, so it is better to record at the lowest possible preamp setting and raise the volume afterwards. Of course that will affect your ability to monitor the recording properly, so try to find the balance between noise and monitoring level.
I think that the idea is to pinpoint the different types of noise and choose which to avoid depending on the tools used. There is the self noise of the microphone, there is the noise caused by the preamp and there is the noise caused by the A/D conversion. Microphones have a constant S/N ratio, so raising the preamp value will raise their noise accordingly. Preamp's self noise is raised exponentially as you reach higher gain values. And finally A/D conversion noise is unchanged no matter the preamp gain and microphone. So recording with low gain VS recording with high gain: The mic S/N is constant in those too cases since it is the ratio of the signal to the mic's self noise. The preamp noise is higher the more gain you add to it so the High Gain recording will have more preamp noise when compared at the same playback level with the Low Gain Recording. And finally the A/D noise floor will be further away from the signal in the high gain recording, when in the Low Gain it will be raised along with the signal in playback.