How can a 40 watt amp outshine a 140 watt amp

Schipo, I don’t know why people don’t pass Clark’s test. But I do know that tests can be constructed to conclude in misleading results. Heck, forget about amplifiers sounding different. What about tube rolling. I hear remarkable differences from same value tubes.

I have an idea about what is going on w the Clark test. I am not a statistician but I do not think that you need 2 sets of 12 out of 12 correct answers to reject the null hypothesis that the 2 amplifiers sound the same. I would that that the chance of getting even one set of 12 100% correct is pretty low, esp. when you have audio memory/human perception issues. Its been a long time since I thought about/learned experimental design but I think what we are talking about here is the difference between a false positive (hearing a difference when there is none) and a false negative (not hearing a difference when there is one). One of them is called type 1 error and the other is type 2 error. I'll bet that if someone here knows a statistician or experimental design specialist, they could work out the numbers pretty easily. I'd be surprised if the ability to get 12/12 correct, twice in a row is pretty darn remote.

02-20-11: Schipo
Richard Clark normally has CD source, amplifiers, high quality home audio speakers, and listening environment set up in advance. But if the listener requests, they can substitute whatever source, source material, amplifiers, speakers (even headphones), and listening environment they prefer, within stipulated practical limits.

Does this mean that Richard Clark will set up the test in my own listening room, with my own equipment? I presume the answer is no.

And that is what is wrong with the test.

I believe that the ABX nature of the test is an illusion. To call it an ABX test is to say, among other things, that only one variable changes. That is of course true in one respect, namely that the amp is the only component that is switched during the test. Hence the test appears to involve a single variable change.

But there is another respect in which hundreds of variables have changed the moment you sit down to take the test, namely the TESTING SYSTEM ITSELF is different from your own. That is the reason, I suspect, why no one can pass the test.

By conducting the test with a system that a participant is not extensively familiar with, the participant’s auditory frame of reference is eliminated. Without it, detecting the manipulation of a single variable change is hopeless.

An analogy: If you put a dish I’ve never eaten in front of me and ask me to tell you if it has ingredient X, I may not be able to do so with any reliability greater than chance. That may be true even if I know what ingredient X tastes like. But if the dish is one my wife has cooked once a week for three years, I can instantly tell you if it has a different ingredient. The reason: Familiarity with the dish.

That’s what missing from Richard Clark’s test: familiarity with the testing system (including the listening room).

Bringing one or two familiar components to the test isn’t enough to make the testing system truly familiar, since literally hundreds of variables are still new to the participant, so Richard Clark's accommodations to participants gives the appearance of scientific rigor without actually providing it.

Bryon

Swampwalker: your making sense when you state 12/12 correct, twice in a row is pretty darn remote.I would gather if Clark lowered the numbers 12/12 correct,there would be a better chance of passing the test. The more you play and listen the more confused audio memory/human perception becomes?

Phaelon: I don't know myself but they all seem to fail...I don't think the testing was constructed to mislead but to teach. I remember Arthur Salvatore writing somewhere when he opened a very expensive mono amp, asking why so much for so little in parts.Maybe the ride is more expensive when it comes to high-end? And only deep pockets need apply.