Effective?
1) Accelerated life-cycle testing in a lab.
2) Statistical analysis based upon a large and diverse study group.
Neither is cheap or easy.
How about a half-arsed test method! That'd be easier. Sorry, just being cynical.
For the accelerated life cycle testing, you could use high temperatures to increase the precipitation effects at the expense of testing for biological growths (high temps kill off critters). So, you could do batches of coolant that were "cooked" for different lengths of time and then test their effects upon pumps and then test for their effectiveness at continuing to provide protection against biological growths and galvanic corrosion.
There's always the chalkboard research method, where you look at the ingredient list and just estimate what long-term effects will be and then make a wild guess as to the magnitude of problems you will encounter.
Fri 18 Mar 2005: edited for spelling and forgetting corrosion protection