Head-Fi.org › Forums › Equipment Forums › Sound Science › Testing audiophile claims and myths
New Posts  All Forums:Forum Nav:

# Testing audiophile claims and myths - Page 102

http://en.wikipedia.org/wiki/Argument_from_ignorance

These tests are to determine the existence of some kind of effect (thing, whatever), that it is nonzero.  The manner of proving things is by providing evidence of counterexamples, to prove that something is false.  You assume that the effect is zero, and test that idea.  (So why not assume that the effect is 1?  Then why not 0.8?  -2? 1203123? pi?  That's why we assume it's 0.)

Based on any experiment, any data collected, you get some kind of evidence regarding this effect.  From that, you can determine the likelihood of there having been some kind of effect.  If the likelihood is high, then we say we have enough evidence to say that we have proven (to a certain degree of certainty) that the effect is nonzero.  There seems to be some kind of deviation as a result of the effect, even if you don't know with certainty what that might be.  You can do further analysis to quantify the nature of the effect, with some confidence levels, and other properties.  Regardless, the explanation of "sure, there is some difference" is confirmed, with a level of confidence.

If there is not enough evidence to say that (most likely) the effect is nonzero, then that is a weaker result.  That means, within some level of confidence, we think that the result could be 0, it could be nonzero like 0.1, maybe nonzero like -0.2:  anyway, we think the range includes 0.  So we're not sure.  We failed to prove that the effect was nonzero, and that's it.  In other words, we don't have enough evidence to distinguish between the possibility between some nonzero effect (say some weak effect) and no effect.  Also, there can be a number of other explanations for why there is not enough evidence.

Anyway, if the underlying theory and other results (prior to experimentation) suggests that there is no big difference, and nobody's been able to demonstrate evidence that a difference exists, then that's that.  We wait for evidence to show up to prove that it exists, and in the meantime, we remain skeptical.  There are bigger things to worry about than the possible existence of giant stealth-camouflaged chipmunks flying through the sky that communicate telepathically across space-time to their overlords in Atlantis.

Quote:
Originally Posted by mikeaj

There are bigger things to worry about than the possible existence of giant stealth-camouflaged chipmunks flying through the sky that communicate telepathically across space-time to their overlords in Atlantis.

You're completely underestimating just how giant these chipmunks are.

Nice explanation though.

JadeEast, while you bring up an interesting point, I think you might be getting too wrapped up in the semantics and might be misinterpreting them. What you're arguing against is a well established thing, and there are very likely already responses to what you have said. I personally don't have any more than what I've already said but I'm gonna look some stuff up later since you've got me thinking.

EDIT: Well, as far as I can tell you are correct JadeEast. The statement appears to be a bit of a paradox. I was under the impression it was the basis of the burden of proof but I don't think it actually is as well established of a statement than I thought it was.

Edited by chewy4 - 10/18/12 at 2:49pm

@ chewy, I wonder if my original example would have been seen differently if I had used the same logical structure but used a different example.

(yes, this is a gross simplification of the issue)

If P then Q.     If (cables make an audible difference) then (blind listening tests will show a statistical difference)

Not Q.            Blind listening tests don't show a statistical difference

Thus Not P.    So cables don't make an audible difference

As far as  "you can't prove a negative" goes, even if it is meant to refer to the non-esitence of something, I find it hard to believe that there would be any arguments against my saying "no live elephants currently exist in my refrigerator."

Still some problems regard absolutes though. There is an extremely unprobable chance that the people doing the blind tests were too hasty, didn't take them seriously etc... and that the elephants are invisible, there are miniature black holes that are absorbing the light reflecting off of the elephants making them impossible to see etc...

To support your case, a good example would be:
In order to be a bachelor by definition, one cannot be married. Therefor, married bachelors do not exist.

Can we assume that all statements made here include the usual disclaimer of, "within the bounds of scientific uncertainty."

You are, obviously, correct. But I don't think we really need to spell out that the likelyhood of either elephants having adapted to invisibility (and size constraints) or miniature black holes, is so extraordinarily small, that, in the absence of any evidence to the support them, they do not merit specific mention.

Honestly - it feels like any moment this thread is going to head towards the metaphysical. Bertrand Russel's cosmic teapot is a nice answer for this entire issue. Lets stick to science where we can. I don't need to prove a negative, if the assertion has no positive support, the default answer in the absence of evidence, is the null hypothesis, or at best a "all the evidence we have to date suggests X". The end.

Edited by liamstrain - 10/18/12 at 4:06pm

You can logically prove negatives, go through all the steps of ifs and thens.  That's not relevant to the original topic.

There are a few issues here, when we're talking about experimental results:

1. There could be issues with the way the experiment is run resulting in bad data (e.g. results skewed from some variable not controlled properly)
2. The design of the experiment is not sufficiently good—not the right participants, not high enough sample size, test not sensitive enough, not testing for particularly the right thing
3. Even assuming there are no problems with the results, the issue of rejecting the null vs. concluding the null is true (properly, the null is rejected, but it may not be true)

With an experiment, you can't prove that the null (hypothesis: no difference) is true.  See my last post.

However, with some null results, assuming #1-2 are covered, you can establish that the likelihood that there is a big difference, is small (even smaller).  It's useful to make a comparison with the kinds of effects that have been demonstrated to be audible, to get a perspective on things.  If some effects are so small that nobody can seem to prove that they are audible, are they worth pursuing?  That's a value judgement.

I agree about the cosmic teapot.  That's really what's at stake, though many assertions are of course a little less ridiculous.

Quote:
Originally Posted by liamstrain

Can we assume that all statements made here include the usual disclaimer of, "within the bounds of scientific uncertainty."

You are, obviously, correct. But I don't think we really need to spell out that the likelyhood of either elephants having adapted to invisibility (and size constraints) or miniature black holes, is so extraordinarily small, that, in the absence of any evidence to the support them, they do not merit specific mention.

Honestly - it feels like any moment this thread is going to head towards the metaphysical. Bertrand Russel's cosmic teapot is a nice answer for this entire issue. Lets stick to science where we can. I don't need to prove a negative, if the assertion has no positive support, the default answer in the absence of evidence, is the null hypothesis, or at best a "all the evidence we have to date suggests X". The end.

Yeah, always a good idea to make that assumption. I do of course agree with the argument in that case(assuming tests were performed with proper scientific standards). Just stating that when making statements about absolutes, you have to consider absolutely everything. But this discussion of logic has gotten a little far off topic at this point

But yes the bolded point was the general idea of the statement that spawned all this.

Edited by chewy4 - 10/18/12 at 4:25pm

Now that we have that solved, what is the meaning of life?

Quote:
Originally Posted by bigshot

Now that we have that solved, what is the meaning of life?

To give life meaning. Next.

Edited by xnor - 10/18/12 at 4:47pm

Probably some hogwash about the human spirit.

Quote:
Originally Posted by bigshot

Now that we have that solved, what is the meaning of life?

Define life.

Then define meaning.

I'll come back later. :)

Life is what you get if you commit a horrible crime.
Quote:
Originally Posted by chewy4

Probably some hogwash about the human spirit.

hahaha you guys crack me up. but seriously, since we are so susceptible to placebo, does anybody feel any envy towards those who do experience some sort of psycho-acoustic effect when switching cables, etc? sure, their minds are playing tricks on them, but doesn't it sound more....fun? i do have hazy recollections of such effects when i was a long-haired Jerry-loving youth and psilocybins were in vogue...

Quote:
Originally Posted by doublea71

hahaha you guys crack me up. but seriously, since we are so susceptible to placebo, does anybody feel any envy towards those who do experience some sort of psycho-acoustic effect when switching cables, etc? sure, their minds are playing tricks on them, but doesn't it sound more....fun? i do have hazy recollections of such effects when i was a long-haired Jerry-loving youth and psilocybins were in vogue...

I don't envy those who pay for placebo. When it comes to placebo, I find .flac to be a very good source of it for me. Even though I know I failed a DBT, I still use it because I "feel" an improvement in SQ when I'm using it. Using .flac, unlike silver cables, doesn't cost me any money.

:shrugs:

New Posts  All Forums:Forum Nav:
Return Home
Back to Forum: Sound Science
• Testing audiophile claims and myths