That doesn't preclude changing that variable in different ways.
Example of my suggestion set up as an experiment:
Control group: cartridge contact kept in dry conditions
Experimental group 1: contact blown on
Experimental group 2: contact kept in moist conditions (moist basement, under dirt, etc. - this could be helpful for finding out how bad blowing on something is compared to something else)
What is important is having an objective way of testing the contacts. If you're testing insertions of a dry cartridge you can't pop it into a NES that's just had a wet cartridge removed.
My guess is that some error or bias could creep into the results, or be artificially removed. An NES that has had moist cartridges inserted regularly will probably start to have a messed-up connector, and while I would carefully dry it to test the effects on only the cartridge connector, you see that this is really a change from the actual real-world situation. In the closest test to realistic you'd have to let both the system and the cartridge dry out normally.
To make sure the cartridges are somewhat equal you'd have to find some brand-new games to test this myth out, or do it over a whole bunch of different cartridges so you can average it out and lessen the effect of wear and tear - they'd have to be the same game, and ideally they'd each be carefully tested by the start for reliability - inserted so many times each before you start testing so that obvious problem cartridges are removed from the equation.
Doing a test with just two cartridges will prove nothing, since we can attribute all the failures to some unseen and uncontrolled variable interfering.