[GrimJosef]

Thanks Ted, I will try to locate those 1978 HFNs - they obviously contain some really interesting 'real time' debates. One of them will also have Barry Fox's write-up in it, I guess. It sounds like he agrees with Moir about the outcome. Moir applied two statistical tests to the data from the listening panel and says that the outcome was "that the residual preferences expressed by the panel were no more than would be achieved by sheer chance (guesswork is the crude term)".

I've rechecked the results of the Colloms tests and it turns out I hadn't remembered them accurately. In fact, of his panel's 13 members 11 could not rank the three amplifiers they were testing on a statistically significant basis. Two could however, and furthermore they agreed on which one was the worst, claiming that the other two were very similar. In another part of the test the listeners were played the same piece of music twice. Sometimes the amp was changed between the playings and sometimes it wasn't. Each listener was asked whether he/she could hear a difference. Apprently most of the panellists did score better than chance on this test.

Cheers,

GJ