It also needs to be conducted by a panel of listeners NOT without familiarity bias. The listeners should be able to repeat these listening sessions over a course of several days to see how their opinions change over time.
Seriously how crazy do we really have to get with listening tests. Any test that doesn't disclose biases (sighted or blind) should be suspect.
There are obvious differences in how loudspeakers sound. No comparison test is perfect, but I can tell which speaker I prefer the sound of in both sighted and blind tests, especially if I am given enough time to evaluate them in the comforts of my own home. I am sure most people can but of course promoting "Blind tests" as the only valid comparative tests is yet another great marketing tool to exploit.