What I tend to do is make a stack of heavy coins, often British ones, which are super weighty (9.5 g for the pound, 12 g for the 2 pound coin), and determine the highest amount of those under which the switches definitely DON'T actuate. Then I use smaller coins, like 10 p, 20 p, etc. and combinations of those, to narrow it down as far as possible. Using heavy coins has the advantage that you end up with a very compact stack, which helps with keeping it straight and upright, both of which influence the results. I test every switch a whole bunch of times, and after they bottom out, I lift the coins as little as possible as is necessary to reset them, and then very very carefully release the coins on the plunger (never on keycaps if at all possible! Fortunately this is a doddle with Alps sliders) without imparting any impact force (essentially I try my hardest not to actuate them). Then I do the same on a a bunch of other switches from the same board, preferably a NOS one. I note down the combinations of coins necessary to actuate the switches, and bring those exact coins to work and weigh them on our analytical lab balance. For amber Alps, for example, which I weighed today as a matter of fact, I found that NONE of the switches actuated at 73.4723 g, but ALL of them actuated at 75.4456 g, so I suspect they're weighted at roughly 75 gf.