[
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2008, Audio Asylum, all rights reserved ]
[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]
87.185.74.71
In Reply to: RE: There are two cases you need to consider: posted by Jim Austin on March 29, 2008 at 06:02:26
>> However, you can also test the hypothesis: are the results drawn from a population
>> that cannot hear a difference (i.e. a random population)? If the answer to that is
>> yes then you have shown to a given level of confidence that a sound cannot be
>> perceived.<<
>
> 1. The test is population based. Then what you write above is true as a sort of
> population average--or, to put it another way, it's completely true IF you assume a
> uniform population. Add individual variation and you get a very different conclusion:
> If one individual, or a subgroup, can hear the difference, then it exists (but
> usually that's not what's being tested). Even in clinical trials of pharmaceuticals
> that aren't approved, some people always benefit.
Population refers to the full set from which the samples in the experiment were drawn and is a fundamental concept in statistical analysis. Identifying the population is part of the definition of whatever the experiment is measuring. It may be multiple samples from a single individual, single samples from multiple individuals or whatever.
If some of the samples are from people that can hear a difference then the distribution will not be random and the hypothesis will not hold.
> 2. If the test is administered to an individual, so that there's truly no intra-
> population variation, there's still the issue of sensitivity, which falls with the
> test duration (correlated with fatigue, apparently). In this case my earlier
> statement applies: that a null result fails to establish perception but does NOT rule
> out the possibility that it might be perceived (so you can't prove that it's NOT
> real).
I think you may be struggling with what an experiment, measurement and sound perception means within the scientific method. Sound perception is not constant and varies with a range factors. This means that the "form" of sound perception is tied to the experiment and different conditions may well produce different results if near a threshold. But what is measured directly in typical audibility experiments is sound perception. Sensitivity is completely meaningless. For sensitivity to have some meaning the experiment would have to measure an indirect quantity that is related to that which the experiment would like to measure but is unable to do so in a direct manner.
> In either case the conclusion is the same: you cannot use an ABX test to rigorously
> conclude that there's no effect on the sound (thus on the possibility of perception).
Since neither of your points is correct/relevant, I am afraid this does not follow.
> ABX testing can be used to establish an effect (so your first claim in the above
> message is true) but cannot be used to rigorously reject a hypothesis...so the second
> claim in your above message is false. It's an important distinction.
We would seem to be back to simple logic again. There are two separate hypotheses and both are being tested in the same manner for acceptance not rejection. It is not difficult to understand making your responses somewhat suspicious.
Surely you must realize the desperation/dishonesty in your attempts to make your beliefs about cables fit the scientific method? Why don't you simply take the view that science is wrong and audiophile beliefs about cables and sound perception are right? It would seem a much more honest position to take.
Follow Ups:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2008, Audio Asylum, all rights reserved ]