ezkcdude Wrote: > Do you do at least 16 ABX trials? That's about the minimum that can give > statistically signficant results.
I'd really like to know who started this "rule of thumb", because it's wrong. If you can tell the difference between A and B 100% of the time, you'd only need 5 trials to get statistical signifance beyond the 0.05 level. This is because the calculation of an exact p-value under the usual null hypothesis is simply 0.5^5 = 0.03125, which is less than 0.05. On the other hand, if the difference between A and B is so slight that you can only tell the difference 51% of the time, then you're going to have to do a lot more than 16 trials to detect that difference. -- Mike Anderson 'FREE RADICAL RADIO!' (http://nvo.com/cd) Hours of free radical MP3s. ------------------------------------------------------------------------ Mike Anderson's Profile: http://forums.slimdevices.com/member.php?userid=1705 View this thread: http://forums.slimdevices.com/showthread.php?t=19818 _______________________________________________ audiophiles mailing list audiophiles@lists.slimdevices.com http://lists.slimdevices.com/lists/listinfo/audiophiles
