The only point where either voltage or current affects the accuracy (of an electromechanical type) is where they get so high they fry the meter or saturate the cores, the disk is moving fast enough that it's at an appreciable fraction of the speed of the moving field, or the total power is so low that friction dominates. The meter doesn't care whether a low current*voltage product comes from low current, low voltage, or both.
But I expect frequency to be a major issue.
Assume, for the moment, that it's low enough that the fields penetrate the disk nicely. Also assume that the moving field from the coils is the same strength as the damping field from the magnet. The disk should stabilize at the speed where they produce equal but opposite drag - which means at half the speed of the moving field. But the higher the frequency, the faster the field moves. So this speed is directly proportional to frequency.
So the meter's reading should be the product of the frequency and the power. OOPS!
This means at 30 hz it will read half the power that flows through it, at 15 hz a quarter of it, and so on.