Quote from: wiinippongamer on September 10, 2012, 05:26:46 AM

@anedote Placebo and "Expectation bias" are the same exact thing...

No, it's just that few recognize the importance of the distinction and this distinction is usually glossed over for convenience. "Placebo" and "nocebo" are categories of anomalous results in the testing of medical treatments which aren't limited to psychological attitudes as befitting of their common usage in audiophile debates, they may result in physiological changes, real cures. The reasons behind placebo effects are irreducible to expectation bias alone. Some such results can even be the complete opposite of what is predicted by expectation bias (in nocebo). Moreover studies which take lengths to eliminate expectation bias by informing a test group that they are receiving a sham treatment still produce placebo effects.

Suffice to say that although there's plenty of evidence that expectation plays a large role in the placebo effect in medicine, we're talking about judging audio products, and it's insulting to trot out the same mysterious mechanism used to describe a cancer remission after the administration of a sugar pill as a synonym for expectation bias because of exposure to advertisements for a brand of cables.

Thanks for the distinction. Makes perfectly sense 

I don't think learning to ignore these factors is impossible. I've had countless situations where I've thought "I just don't hear any difference" or "this sounds different but I don't think it's better" or just "this is crap". Sometimes it's very close one way or the other, and in those situations blind testing can be useful but I wouldn't say it's necessary in all cases. The only way I feel blind testing is useful though is when the method is exactly the same as what one uses for sighted tests - same songs, same listening time, same system etc.

I think blind testing of large groups in unfamiliar locations with unfamiliar systems with unfamiliar songs played at pre-determined durations is utterly useless. That might be good enough to tell a Maggie apart from a B&W, but for two largely similar DACs? Forget it, you'll get 50/50.

Blind testing is an essential tool in accounting for expectation bias, but it's not the be-all and end-all. I'm all for the listen and buy whatever tickles your fancy philosophy. It's a sort of healthy hedonism that enables us to enjoy music. Tests and measurements must come in at the market sense. Unfortunately few headphone manufacturers take the time to include DBT in their rush to market, so a greater push is necessary to educate consumers and raise the level of criticism.

Take for instance the limits of human hearing. I'd like to see a breakdown on the percentage of people run across figures of 5-500khz bandwidth who know that studies place the realms of audibility with the limits at no more than 15-23khz. At that point I think it's fair to the customer to know that they're being toyed with.

Just because I disagree with someone on a few issues doesn't mean I can't agree with them or respect them for their work in other areas.  I mean, I think purrin is dead wrong in a lot of the stuff he says about amps and DACs but I think his CSD plots are an amazing advance in headphone measurements.  A few other people do them with headphones but for the most part their techniques are flawed and the results are too inconsistent to be very useful.

LOL. So funny to read this here. I posted almost the same thing a few days ago on Headfi saying I preferred the sound of amps without huge amounts of negative feedback and got called out on it by a by the book measurements guy 

Someone posted this on another forum, but I think Charles Hansen (and Nelson Pass) do it just right. Charles does no NFB amps, with very little local NFB that sound great and measure good. My educated guess is the slightly higher output Z on the MX-R monos I like gives it a touch less of that stereotypical clinical SS sound.

That's the hot topic. I've read some of Pass's articles from back in the day, although I lack the requisite knowledge to fully understand them. From what I've gleaned from celebrity engineers (& amp; celebrity speaker makers) is that they primarily deal with a sample size of the only ones they trust - themselves. I can't fault the technical arguments behind reducing crossover distortion in amps, but I'd like to call for real testing to see if there's a noticeable difference to a random set of testers who don't give a fuck about circuit design.

I wish that engineers, manufacturers and audio enthusiasts would glean more from the sort of testing methodologies employed in epidemiology. Ironically this would put the placebo effect in the forefront, so I think it wiser to test for specific biases and experiment with cohort studies (say, head-fi members vs. general population). I recall there's a head-fier who is by day an analyst for medical equipment testing. He is sometimes active on sound science and mentioned that he wanted to employ a robust study design during one of the meets. If there's an opportunity at something like H3 next year, or at a canjam, I'd be glad to fund an effort to get 20-30+ people to spend an hour in the DBT funhouse.

anetode, I basically agree with your thoughts about the subject.

Sometimes people are just too closed minded to understand that during certain times, AB testing is not always the way to know whether something is better than something else. Blind testing has its merits though, and I would even accept a cable challenge to tell which one is which but only if I am very familiar with all the other factors like the music, the headphones used, and all other parts of the chain. Also, even though I may be able to tell which cable is which, I still may not be able to say which cable will sound better to me in the long term right then and there. Something like this just happened to me recently by the way so the issue is still fresh in my mind

By the way something else about placebo...

I am now looking at that question on innerfidelity to Purrin about the measurement method he uses. I believe that in itself will be placebo as to whether people will believe if Purrin's methods are good or not even though the people asking the questions don't listen to the headphones in question and compare it to the graphs first. Heck, some of these guys might not even know what a graph is for except to see which heaphone looks flatter without understanding how they should also give weight to the relative aspects compared to what they hear.