Interesting article. I'm a big fan of blind tests for anything that is influenced by unrelated perceptions.
My biggest quibble with this test has to do with using music unfamiliar to the users to evaluate the speakers. I'd suggest something along the lines of creating a compilation CD of all the tracks you're using for evaluation, sending it to participants and having them listen to all the tracks multiple times on their own sound system. That way they'd be familiar with the music and how it sounds on their reference system, which would free them to pay attention to how the speakers reproduce the music.
