What I do is apply a higher voltage than the regulation voltage with a resistor in series to limit the current, then check the regulator voltage with a DMM to make sure it's in spec. Two 9V batteries in series (~18V) is great for a 12V regulator; throw in a 1kΞ© resistor in series between the battery and the positive terminal on the regulator and you should be in business. You won't want to use any random resistor here, too small and it will burn up, too large and you won't have enough current to turn on the regulator. A 1kΞ© is a good balance for most diode based regulators (36mA in this case).
In theory a 6V bulb would work in place of that resistor, but I haven't tried it. It should light up with a good regulator. I'd use a low wattage one, like a little 1W speedo light or something. 9V batteries are pretty wimpy.
If the regulator is bad, or if you aren't giving it enough current to turn on, the voltage across it will read the full 18V. If it works and the test setup is correct, it should read about 12V. Many "12V" regulators are really designed for charging 12V batteries, and can regulate at voltages up to about 14.5V.
As you know, moped magnetos don't produce true AC power, so DMM voltage readings on a running bike aren't very useful. The scaling factor between what the DMM says and the peak voltage varies by configuration; how many coils, how physically large those coils are, how many magnets, on the flywheel, how large the magnets are, etc.
This is also complicated by the fact that there are many ways to build a regulator. A simple zener diode regulator can be checked with a DMM in diode mode, but who knows if that's what's going on inside these mysterious lumps of epoxy.