jeffmeh wrote: > P Floding;191291 Wrote: >> Let's just stress that a negative AB or ABX means very little. >> Certainly proves nothing in the general sense. > > Agreed, but let's also stipulate that it does prove that the SPECIFIC > subject of the test CURRENTLY cannot discern an audible difference. > Note the emphasis, as it says nothing about other subjects or even that > subject in the future.
Statistical testing never says anything about specific people or samples, and only addresses probabilities in general. Well, we could say that at the time of the test, that specific person found X. But it says nothing about whether they would fix X again if tested again, or if they would find Y. What random testing does is expect that people would in general be 'right' 50% of the time, just like guessing a fair coin flip. That you test and find 10 heads in a row does not mean that the coin was rigged (unfair) there is a non-zero probability that will happen in a fair random test. The proper phrasing is to say "the null hypothesis is that there is no difference detectable by listening to a SB3 versus a SB3 feeding a Larvy" and then say that in our testing, we reject the null hypothesis with a confidence level of 95% (or 99% or whatever you want). Meaning that we are 95% sure that there is an audible difference. Difference does not always mean better. -- Pat http://www.pfarrell.com/music/slimserver/slimsoftware.html _______________________________________________ audiophiles mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/audiophiles
