http://www.qscaudio.com/support/library/manuals/abxman.pdf
Page 10 has a nice and simple explanation of the statistics. 8 out of 10 would certainly satisfy me (and most other skeptics) that there's something going on. 9 or 10 out of 10 would get the trial publishable.
Page 10 has a nice and simple explanation of the statistics. 8 out of 10 would certainly satisfy me (and most other skeptics) that there's something going on. 9 or 10 out of 10 would get the trial publishable.
The phrase "about a 95% confidence level" is right as it means ~95% doesn´t it?
The "exact" number for 8 out of ten is 0.05468.... so that fits.
I don´t know what did happen during the calculation of the mentioned table, but would guess that they did calculate for a two sided test and forgot to mention it.
Together with the underlying phrase
"The table on this page lists a recommended range of trials, and the minimum number of correct responses
necessary to reach a 95% or better level of confidence."
the content of the table is wrong, as for example a 11 trial run gives p=0.113 for random guessing 8 or more correct answers.
It might be useful to do two sided tests (Leventhal had a paper on a topic
"statistical significant poor performance") but routinely it is a one sided test.
Wishes
future test methods?
Since at least one "believer" has proposed all sorts of metallurgical, anisotropic micro-rheological like, and surface anisotropies, not to mention dielectric and insulative issues, with no corroborative evidence of any sort, , as probable causes for "burn in" and other cable differential behaviors, perhaps a Plackett-Burman type multi-factor design could yield some insight into the reality of said claims and beliefs and put them to rest
Not that I'd be interested in setting such an experiment up;
, but it could provide an efficient screening methodology, assuming 2nd and higher order confounding has minimal effect. If not the case, there are other Hadamard matix type designs which significantly minimize trials and maximize results. The fly in the ointment is doing the up front work to design the tests properly.
A brief summary at the Wiki article:
Plackett–Burman design - Wikipedia, the free encyclopedia
Since at least one "believer" has proposed all sorts of metallurgical, anisotropic micro-rheological like, and surface anisotropies, not to mention dielectric and insulative issues, with no corroborative evidence of any sort, , as probable causes for "burn in" and other cable differential behaviors, perhaps a Plackett-Burman type multi-factor design could yield some insight into the reality of said claims and beliefs and put them to rest
Not that I'd be interested in setting such an experiment up;
, but it could provide an efficient screening methodology, assuming 2nd and higher order confounding has minimal effect. If not the case, there are other Hadamard matix type designs which significantly minimize trials and maximize results. The fly in the ointment is doing the up front work to design the tests properly. A brief summary at the Wiki article:
Plackett–Burman design - Wikipedia, the free encyclopedia
@ SY -
Without me having to go back and re-reading 20 pages again... Is there anything preventing a test-ee from making measurements and discerning a difference that way? - Essentially cheating by not using their listening skills alone?
I'm not as knowledgeable as you are regarding the ability to use instruments to 'measure' a burn-in difference. However, if you are sure that there is no measurable difference, I'm fine with it.
I do know that I have made measurements between components before which show a plainly obvious difference, even though no one could prove they heard anything.
Also, have we defined what will be used to listen? Is it music? Is it white / pink noise? Not that anything will make a difference, but again - unforeseen consequences.
Hey Oatmeal - this thread had been quiet for over 24 hours. I thought SY had killed it off.
From my reading of the test, there is no reason the person being testing can not measure. It is not believed that the differences will be measurable. If they are, that's certainly interesting news.
EDIT. SY posted at the same time.
From my reading of the test, there is no reason the person being testing can not measure. It is not believed that the differences will be measurable. If they are, that's certainly interesting news.
EDIT. SY posted at the same time.
So you should be. You get some nice new cables out of this, whatever the outcome...
So you should be. You get some nice new cables out of this, whatever the outcome...
I guess jleaman would want his cables back, if they are better than mine, perhaps I will send them back one at a time.
Are we moving forward with this?
Logistics: should we take this offline? PM, email? Let's hear again from those who can commit.
1) At the least we need two cables.
2) The person who's got the cables will have to give us an estimate of shipping the two cables to the burner, so that we could get the money together and transfer it.
Logistics: should we take this offline? PM, email? Let's hear again from those who can commit.
1) At the least we need two cables.
2) The person who's got the cables will have to give us an estimate of shipping the two cables to the burner, so that we could get the money together and transfer it.
I think that if you guys are going to seriously test the idea of speaker cable burn in that you should make sure (1) There are no Shatki stones in the area- not even in the house AND (2) No CD demagnetizers in the room AND (3) there are no power line conditioners connected to the burn in equipment used in the experiment.
I read a thread once where all these things were co-existent and a bright light from a corner in the room slowly emerged during the cable burn in time, the experimenters had to run out and shut the house power off. It has never been determined what the source of the light was, but some speculate that it was Jesus or Faradays ghost.
Any two of the above things is probably harmless though.
I read a thread once where all these things were co-existent and a bright light from a corner in the room slowly emerged during the cable burn in time, the experimenters had to run out and shut the house power off. It has never been determined what the source of the light was, but some speculate that it was Jesus or Faradays ghost.
Any two of the above things is probably harmless though.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.