As you said in the question yourself, differential signals are usually measured using an oscilloscope. Using a multimeter may only give you reasonable results for DC-Offsets (or DC common mode signals).
Using a multimeter to measure the "changing" part of the waveform will generally fail due to several reasons:
- Most multimeters have bandwidths for RMS measurements up to 10kHz. Differential signals are most often used way beyond that bandwidth.
- Multimeters assume sinusoidal waveforms. If you were measuring a pure rectangle you could work with a conversion factor, but this will fail for transient signals like digital streams.
- Generally you have very poor control over the aspect of the signal you are measuring, especially when there are "mute" intervals between frames. You will most likely compare two different signals, even when using two multimeters at the same time to measure N and P waveforms, since they are not synchronized.
If you still want to go ahead you should at least perform an analog differential-to-single-ended conversion (using a difference amplifier), so that you are at least not comparing different sections of the N and P signals. If the multimeter inputs are floating, you can also measure directly across the P and N lines. Similarly, you could use an OpAmp adder to measure the common mode part of your differential signal.
To sum up: Unless you specify exactly the multimeter you are using and exactly what waveform you are measuring, the results will be meaningless. Especially when the multimeter is in some kind of auto mode.