What's your view on blind testing?


Double blind testing (DBT) has been a hot topic among engineers and audiophiles for God knows how long. The 'scientific' group believes that a properly conducted blind test will reveal negligible differences amongst audio components (amps, cables, dacs, etc.) while the other group believes that there are noticeable audible differences across manufacturer's in various price points. I know there are quite a few EEs on the board here and I'd like to get your perspective. Thanks for any insight.
gfcf424892
Post removed 
Extravaganza wrote,

"very bad,because normally short memory of human brain remembers only fraction of information which was played.And long term memory remembers only those moments which impressed very much during listening,lets say tonaly,imaging."

Oh, sure, if we're talking about the man under the bridge you might be right, he probably has no idea what he"s listening to. But most experienced audiophiles have excellent short term and long term memory. I suspect the problem with DBT as far as it meaning anything for audio is the idea that a subject must pass 10 trials in a row with no misses. For that reason, I'm out.
Double blind testing is a complex process because neither the test subjects nor the testers are privy to the identity of the materials being tested. The goal is to eliminate bias on both sides of the experiment. Even if long term audio DBT could be done with sufficient statistical power (lots of subjects, lots of trial sessions, lots of recorded material, ideally only one type of test component), self reporting questionnaires are notoriously unreliable. DBT methodology is sound when it can be reliably applied, but I seriously doubt that's the case with reviewing and rating audio equipment.
We're doing it again. Letting perfection become the enemy of good. If we can't have perfection we will have nothing and good will not be done.

There is no reason to try to set up a perfect double blind test. How about having someone outside of Stereophile set up a system in the Stereophile offices with say two different amps that are covered and a device to switch between them.

When the reviewers are in the office, they can listen and state which amp they prefer. After most have given their preference, the identity of the amps and who liked which or if they were able to distinguish between the amps would be revealed. This would be done for fun and enlightenment.

There is no need to set up a statistically valid double blind test. Isn't this hobby supposed to be for fun (and enlightenment)? Let's have some fun and let go of our tightly held dogma.