It's there, but doesn't afect the results as long as it's short-term stable. Of course, you *do* have to accept Ohm's Law for the duration.
I assumed that a pic would only digitize ground-referenced voltages, hence the path through the math. When I do this, I usually use a fully differential adc, so I can measure the voltage across each resistor directly, which has the advantage of ignoring any lead resistances.
John