Probing for resistance with signal voltage present can create distortion in itself, as the meter despite its high impedance is a parallel in the case of the series pairs input to output. A basic set up would assess the attachment of measuring equipment by having music playing to see said equipment and its many settings was not the actual distortion contributor, caused by being a parallel.
I think we may be talking slightly past each other here.
Parameters such as resistance, temperature coefficient, and voltage coefficient are all measured on the resistor in isolation. To get the resistance and voltage coefficient, you measure the I/V curve of the resistor. To get the temperature coefficient, you measure the I/V curve at various temperatures. You can see an example of this in the paper I linked to in Post #64. You can then develop a SPICE model that matches the measurements and use it in simulation to determine the performance of any circuit that includes that resistor. A circuit as simple as an LDR-based volume control (two resistors) would be trivial to simulate and as long as the resistor model is reasonably accurate, it's reasonable to expect the simulation results to match reality.
You can measure the resistance with the music playing by measuring the voltage across the resistor, the current through it, and calculating the derivative. That will give you the same result as just measuring the I/V curve, but is much more complicated to set up and replicate, which is why we use source meters and such rather than music players for resistor characterization. Similarly for the AC parameters of the resistor. Swept sine wave vs music will ultimately yield the same parasitics (which can also be included in the simulation model).
Tom