Re: Music and Art
- "Yet the test showed that statistically the amps were indistinguishable."
Did the test really show that? What were the values of alpha and beta, with out those, I as a trained statistician wouldn't know. Do you know the values for that test. If you don't know/didn't find out, how can you say that the test showed anything at all?
JBTWay I have raised this before and been pretty much ignored. If the value of beta is over 0.5 than the test shows the commitment of the tester to a particular outcome, OR their complete ignorance of testing.
In plain English there's a higher than 50% probability that the test not only doesn't show what it purports to, but that the opposite is in fact true for that test.
The test you mention might have had the result that - during the test - no-one could distinguish the amps. That does not mean that the amps would be audibly indistinguishable to anyone. And, you've already stated that they are audibly different. The next test should have bean after the subjects had been told one had a bit more bass, to see if they could reliably identify it.
When I make changes to my system e.g. using a commercial quality teflon capacitor to replace a WIMA MKP in the passive RIAA, the differences were audible and I'm not imagining it. My wife said much the same things and she's not technical at all. She is the reason I changed over to MC cartridges in the late 1970s.
I'd like a response this time, thanks.
- Some people even try to assure that they do enough trails to get beta lower than 0.5, and some also know that beta and n trials are interlinked, so they don't get too tight with p, either.
Far too few IMO&E.
--- In firstname.lastname@example.org, tim-bailey@... wrote:
> "Yet the test showed that statistically the amps were indistinguishable."
> Did the test really show that? ......... "