Agreed, the only way that test can be properly done is when the participants don't know what is playing. If you want it super properly done, not even the tester should know what is playing. I was in that room during one of those demos, I didn't stick around because the results would have very limited value. I like Bill, but he isn't really challenging anyone's preconceived notions in a test like that.
Last year at AXPONA there was a room for a power cable manufacturer (Shunyata I think?) who were doing the same kind of test. They were switching from a stock power cable to their magic power cable for the same recording, and acted as though the difference were night and day. I didn't hear anything different, but some of the audience there agreed that the sound was dramatically improved.
Audiophiles have some of the most active imaginations on Earth!