cliveb wrote: > And the crucial word here is *consistently*. If you run a trial with 20 > people, it is statisically very likely that one of them will "pass the > test". What then needs to be done is to repeat the test and have *the > same person* pass the test again. If they don't, it's just a standard > case of regression to the mean.
Indeed though it totally depends on the test. If the test is designed such that the expectation is that no-one will pass it, the norm is zero and the standard deviation is very very narrow. If any person comes even close to consistently "passing" then the result is statistically significant. I haven't read the detailed data behind this paper, just making observations about statistics here. ------------------------------------------------------------------------ drmatt's Profile: http://forums.slimdevices.com/member.php?userid=59498 View this thread: http://forums.slimdevices.com/showthread.php?t=105816 _______________________________________________ audiophiles mailing list audiophiles@lists.slimdevices.com http://lists.slimdevices.com/mailman/listinfo/audiophiles