Blind testing is a bit over rated and often never done correctly by most people, including the manufacturers that tout it as a necessity. Do you need to do a blind test to determine if a gourmet angus hamburger tastes better than a 99cent McDonalds toe-nail burger? I don't think so. Blind Testing has alot of relevancy in the medical field where you are dealing with minute differences, control groups, and huge statistical sample sizes. Unlike the medical field, there are usually huge differences in how speakers sound which even untrained ears can distinguish quite easily.
I always love the manufacturers that claim you must do a blind test else the results will be biased towards the better looking speakers. Does that mean that their speakers (if they lose) are always the ugliest in the competition? If so, why not make them prettier and then not have to worry about controlled sighted tests anymore
Level matching is critical and you should be able to switch at will while also adjusting listening level of both pairs proportionally. It's a good idea to hear a speaker at various listening levels to test for compression as well as how linear it is at low levels too.