Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Medic. study—placebo tends to lack important effect—relevant to audio? (Read 48361 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Medic. study—placebo tends to lack important effect—relevant to audio?

Disclaimer: This is not a defense of snake oil.

Back in the early 2000's two researchers, Hróbjartsson and Gøtzsche, published two analyses of clinical studies in which placebos had been employed and found that there was no observer-reported improvement in conditions where placebos had been used. By observer-reported they meant objective symptoms that could be measured by third parties, like doctors. This was set in opposition to patient-reported symptoms like pain and nausea, that cannot be objectively measured. For these patient-reported symptons some placebo effects had been found, but they were indistinguishable from reporting bias. Basically the only place where placebo effect was to be found was in the patients' eagerness to please and doctors and in their general politeness.

These studies were criticized on methodological grounds, and, since I'm layman, I just dropped the matter. But recently I've found out that the same researchers did a new analysis in 2010 which addressed these criticisms and reached pretty much the same conclusions. Here's the summary:
Quote
Placebo interventions are often claimed to substantially improve many clinical conditions. However, most reports on effects of placebos are based on unreliable studies that have not randomised patients to placebo or no treatment.

We studied the effect of placebo treatments by reviewing 202 trials comparing placebo treatment with no treatment covering 60 healthcare problems. In general, placebo treatments produced no major health benefits, although on average they had a modest effect on outcomes reported by patients, such as pain. However, the effect on pain varied from large to non-existent, even in well-conducted trials. Variations in the effect of placebo was partly explained by variations in how trials were conducted, the type of placebo used, and whether patients were informed that the trial involved placebo.

And here's a link to the abstract: http://onlinelibrary.wiley.com/doi/10.1002...62808AE1.d01t02

They were talking about placebo effect in Medicine, of course, but we can wonder whether these findings don't also invalidate the metaphorical usage of the expression, placebo effect, in audio matters.

When patients report that their pain has diminished after eating a sugar pill they're just trying to please their doctors, or perhaps they're too polite to say that the treatment was completely ineffective, and prefer to say that there was a little improvement. Might it be the case that some audio journalists, when they fail to notice any difference between audio equipment, refuse to say so out of sheer politeness and a good-hearted desire to preserve the feelings of their patrons and advertisers?

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #1
Ya kinda got things twisted in a knot there and are drawing conclusions 180 degrees out of phase with the paper.

The paper is saying that the placebo leads to reports of subjective changes.  Changes which can't be objectively measured (or at best weakly).  But you ask the (opposite) question of "does this possibly mean that placebo leads to (incorrect) reports of no subjective changes?"

This paper is 110% in support of the "common wisdom" here.  Placebo effects lead to reports of (audio) differences where none can be measured by outside objective observers.

All this paper shows which is contrary to popular belief is that placebo effects (in medicine) might not be leading to objectively measured changes in patient outcome.  NOT that the placebo effect fails to convince the subjective opinions of the placebo "eater".
Creature of habit.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #2
NOT that the placebo effect fails to convince the subjective opinions of the placebo "eater".


I believe the two previous analyses were clearer on this point, but just to avoid confusions, let me quote from the 2010 abstract:
Quote
it is difficult to distinguish patient-reported effects of placebo from biased reporting. The effect on pain varied, even among trials with low risk of bias, from negligible to clinically important.


Read that as an academic euphemism for: Placebo can't be distinguished from biased reporting. I interpreted "biased reporting" as desire to please and general politeness for simplicity's sake. Of course there are other forms of biased reporting. Now, logically, you can't dismiss the possibility of placebo effect really existing in patient-reported symptoms. All you can say is that the evidence doesn't allow you to distinguish that possibility from another, namely that the patients' have a desire to please and are generally polite. Once that's established, which of the following two explanation is more likely based on all that's known about human health:

1) There's a miracle cure based on the power of the mind that sometimes relieves subjetive symptoms.
2) What used to be called placebo effect was really just biased reporting.

I'm obviously sticking with 2). To be clear: this is my interpretation, the analysis doesn't take this last step.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #3
That is not the proper conclusion.  As I said you're 180 out of phase.

You're reading into their words.  They are saying that the objective measurable effect of placebo on the patient can not be clinically distinguished from biased reporting.  That is far from saying it IS biased reporting. 

And of course amongst your two explanations #1 is a straw man nicely set up to be knocked down by your own #2.

The actual, supported, and simplest is #3:  The patient believes they experience something, and report it even though objective measurements don't support their belief.


The irony here, IMHO, is the only biased reporting is coming from you.
Creature of habit.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #4
Let's replace 1) with the following definition from http://wordnetweb.princeton.edu/perl/webwn...acebo%20effect:

1*)S: (n) placebo effect - any effect that seems to be a consequence of administering a placebo; the change is usually beneficial and is assumed result from the person's faith in the treatment or preconceptions about what the experimental drug was supposed to do

Once we've accepted that patient-reported placebo is indistinguishable from biased reporting, which of two is more likely, based on all that's known about human health:

1*) A patient's faith or preconceptions in a treatment or drug can produce beneficial changes in subjective symptoms.
2) What used to be called placebo effect was just biased reporting all along.

I'm still sticking with 2). I repeat: this is not reporting, this is not taken from the paper, it's my own inference.

EDIT: Again, this is not a logically valid inference, which is why the paper didn't make it. All they could claim was that the options were indistinguishable. But I believe that those of us who are not bound by the absolute rigors of Science can be a litte freer with our reasoning. Logically I'm not allowed to claim that Flying Spaghetti Monster doesn't exist, but I still feel justified in believing that it doesn't.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #5
So am I now to suspect all the people who say FLAC sounds inferior to wave are only doing so in order to be polite; even when they come here and do it?

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #6
So am I now to suspect all the people who say FLAC sounds inferior to wave are only doing so in order to be polite; even when they come here and do it?

I'd rather suspect that they already believe that wave is superior "without question, it's obvious, everyone knows it for reasons x, y and z", so that even though they can't hear any difference, they would rather claim that they can, following the classic principle laid on The Emperor's New Clothes. Most of the times they might be really hearing diffrences as a result of an improper comparison (i.e. not volume matched, flawed FLAC compression, and maybe even codec problems).

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #7
Codec problems that result in something that still plays?!?

You're grasping at straws.

In the meantime you've failed to take into account what has been documented and understood about the way in which humans can be biased into hearing things that never existed.  How do you plan on resolving the McGurk effect, for instance?  How about all the failed ABX tests where the person submitting them thought they did indeed validate their claims that there were audible differences?

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #8
You're grasping at straws.

I don't think I have the burden of explaining why they do what they do, just that, whatever the reason, it might not be placebo effect. Those suggestions were meant to be an open-ended list of possibilities, I'm sure technically savvy individuals could add other options.

In the meantime you've failed to take into account what has been documented and understood about the way in which humans can be biased into hearing things that never existed.  How do you plan on resolving the McGurk effect, for instance?

Greynol, if you read my original post carefully you will see that I make no claims about audio placebo, but I ask whether perhaps they could be re-examined in light of developments regarding medical placebo. It's certainly possible that people do hear what they think that they're hearing, but until recently I believed that in some circunstances people could be cured by eating sugar pills. If the last one can be false, don't you think it would be interesting to at least consider that the first might be false as well? And remember that there were a myriad studies "proving" that medical placebo was real, but it turned out that the interpretion of these studies was incorrect.

Quote
How about all the failed ABX tests where the person submitting them thought they did indeed validate their claims that there were audible differences?

But how can we know what was really going on in their heads? All we have to go on is what they say, and if what they say can be wrong in more important matters like health care, might it not be wrong also about audio? Again, this is just an alternative suggestion that whatever makes them make their claims, it might not be placebo effect.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #9
Whatever else may be possible (pink elephants silping tea in orbit around the planet Uranus?) the McGurk effect does in fact show that people can and do hear something that doesn't actually exist, TYVM.

might it not be wrong also about audio?
That you don't know enough about it and are skeptical is hardly enough to warrant tossing a substantial body of research in favor of your false equivalency.

FWIW, many people here who accept that humans can be biased into hearing differences that do not exist do not believe that sugar pills work on pain.  These are not mutually exclusive.  Your failure to recognize this doesn't grant the possibility for you to be right.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #10
might it not be wrong also about audio?
That you don't know enough about it and are skeptical is hardly enough to warrant tossing a substantial body of research in favor of your false equivalency.

FWIW, many people here who accept that humans can be biased into hearing differences that do not exist do not believe that sugar pills work on pain.  These are not mutually exclusive.  Your failure to recognize this shouldn't be used to suggest there is a possibility that you could be right.

But might it be enough to suggest that a friendly discussion take place, among willing participants, in an internet discussion forum, where arguments that support the existence of audio placebo effect might be offered?

And maybe it might also suggest that a new popular expression be adopted for the phenomenon now known as placebo effect in audio.

BTW, I actually found your reference to the McGurk effect interesting so I did some homework. So far as I could see it's different from placebo effect because it relies on an actual visual cue, whereas placebo effect in audio could be based exclusively on beliefs or desires. The McGurk effect remains even if the subjects know what's happening. Does the same happen for placebo effect? (i.e. do audio objectivists still think they hear differences that they know won't be found in an ABX test?)




Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #14
You may look it up.  Please don't ask me to conduct your research for you.

it still remains the case that audio placebo effect disappears during a double-blind trial but the McGurk effect remains. That alone is sufficient for me to conclude that they're not necessarily connected, one might be true and the other false.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #15
ABX does not remove the ability for someone to affix differences that do not actually exist to what is "heard" during testing.  Instead it serves to randomize the occurence of those "differences".

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #16
ABX does not remove the ability for someone to affix differences that do not actually exist to what is "heard" during testing.  Instead it serves to randomize the occurence of those "differences".

But the differences become statistically insignificant so they "disappear" from the conclusion. The different sounds heard in the the McGurk effect don't "disappear" from the conclusion. Even after ramdonization and the application of a double-blind protocol, researchers can find a statistically significant McGurk effect. The contrary is true for the placebo effect in audio.


Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #18
This would be akin to asking the question "If the listener is told that he may be listening to a superior system, does that make it truly sound better to him, or is he just saying that it sounds better to please the tester because he really can't tell the difference"?

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #19
The only implication this study has on the way we test for audible differences is that we have no assurance that the person performing the test is being honest

I've read plenty of discussions where this has been taken into consideration in one way or another, so this is hardly a new wrinkle.

Besides, dishonesty during the test is not going to trend towards a decreased p-value.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #20
it still remains the case that audio placebo effect disappears during a double-blind trial but the McGurk effect remains. That alone is sufficient for me to conclude that they're not necessarily connected, one might be true and the other false.


The interaction known as the McGurk effect survives the double-blind trias, and this fact is verification that there is an interaction between our auditory and visual systems.

If anything, this highlights the need for ABX testing of supposed audio differences.  If our visual system can influence our auditory perception, this suggests that other brain functions may also influence our audio perception.  What we expect to hear can influence how we hear, and what we hear. 

ABX tests provide a good tool for eliminating biases caused by placebo effects.

Prior posts is this thread identified 3 possible placebo effects:

1) "A patient's faith or preconceptions produce a beneficial effect."

Audio analogy: What we expect to hear has an influence on what we actually hear.  For example; our expectations may cause us to focus on a particular sound, instrument, or band of frequencies, and we actually notice things that we had not heard before.

2) "Biased reporting"

Audio analogy: We claim to hear a difference that we do not really hear (for a variety of reasons).

3) "The patient believes they experience something even though objective measurements don't support their belief."

Audio analogy: A is identical to B, but listener becomes convinced that A sounds different than B.


All three of the above placebo effects can be eliminated with a double-blind ABX test.

Nevertheless, it should be noted that the ABX test is not perfect.  Any ABX test will cause some disruption to the listening experience.  It has been established that the McGurk effect can be diminished if the subject is engaged in tactile tasks while listening to the audio and viewing the video.  The distraction caused by the tactile task seems to diminish the McGurk effect.

Similarly, the tactile and visual tasks, and mental stress, associated with an ABX test may produce some minimal reduction in a subject's auditory acuity.  The distractions caused by an ABX test may obscure some small audible differences, but the ABX test is very effective for eliminating placebo biases.  ABX tests should be conducted in a way that minimizes the mental distractions caused by the test itself.
John Siau
Vice President
Benchmark Media Systems, Inc.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #21
This would be akin to asking the question "If the listener is told that he may be listening to a superior system, does that make it truly sound better to him, or is he just saying that it sounds better to please the tester because he really can't tell the difference"?


Page 13, right before the “Strengths and weaknesses” heading:
[blockquote]Larger effects of placebo were also found in the trials that falsely informed patients that the study compared two active treatments with no-treatment.[/blockquote]
So yes, it helps if you can pre-convince the patient.



(And there is no surprise to spot confirmation bias in this thread either, by the way.)

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #22
Page 13, right before the “Strengths and weaknesses” heading:
[blockquote]Larger effects of placebo were also found in the trials that falsely informed patients that the study compared two active treatments with no-treatment.[/blockquote]

Ouch.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #23
This would be akin to asking the question "If the listener is told that he may be listening to a superior system, does that make it truly sound better to him, or is he just saying that it sounds better to please the tester because he really can't tell the difference"?

We presume that we have a proper ABX test so we can conclude that no objective difference exists between A and B. Then we ask the preceding question, but let's leave the listener's motivations open - maybe it's just another instance of naked-emperor-disorder, maybe not, whatever it is, the listener doesn't really perceive a difference but still claims that he does.

From that, isn't your quote a perfectly legitimate question? I was surprised by the generally hostile reaction to the thread because if there's no real difference, not even subjective difference, between the two samples, then at least one of the arguments that subjectivists have used to reject ABXing would become untenable, namely that they can really hear the difference at home but not under "laboratory" conditions.

This would also be consistent with the suspicions voiced by some objectivists that some audiophile journalists and manufacturers are not completelly honest in their claims.

Medic. study—placebo tends to lack important effect—relevant to audio?

Reply #24
Quote
In general, placebo treatments produced no major health benefits, although on average they had a modest effect on outcomes reported by patients...
So, a placebo is an ineffective form of treatment, but some patients perceive an improvement!      Is that surprising?   

Exactly what we see in audio!  Is that surprising?