How Superstitions Really Work
 

I'm really troubled by the mindset that justifies self-deception on grounds that it seems to work, if only because it seems logically inconsistent that someone in denial could adequately judge whether something was working.

I've been reading a lot lately about antidepressants and the studies that seem to indicate they only work because of the placebo effect. Of course, there's a lot of evidence on both sides and I don't want to hijack the thread with a debate over which studies are more valid. What strikes me is the philosophical debate that follows if we postulate that they only work because patients believe they work. Should we stop prescribing the drugs or at least publicize this information, at the risk of destroying the positive effects they have on people who believe in them? I can understand why some would say no, but I just can't get on board with that.

1. A placebo is not a real medicine, but it works in many instances, therefore, it is real. Telling people it's not real may stop it from working, but why do that? Same with superstitions and lucky charms etc. If someone thinks they run faster wearing their lucky socks, then why try to prove them wrong?

2. Do we have the right to force our views on others, even if our views are correct and theirs are demonstrably wrong?

Having the truth available is fine, but I don't know that we should be forcing others to believe it.

What is your evidence a placebo "works"? Also, "works" compared to what?

Sorry. A placebo is a pretend medicine. Sometimes it is used in place of "real" medicine because of the "placebo effect" in which a person who believes they are being treated with real medicine may get better, even though they are taking sugar pills or similar.

Every clinical trial has to test a control group, a placebo group and the trial group to get clear evidence that the new medication actually works. It is thought that around 1 in 3 people will feel the effects of fake medicine.

Yes, I know what a placebo is.

I was wondering why you think there is evidence that a placebo works. So I asked you: A placebo works compared to what? If you're comparing a medicine to a placebo are you claiming that the 1/3 of the people get better in that case because of the placebo? So, for example, a placebo is put on wounds in one group, a drug is put on wounds in another. Both groups get better and you're saying that the placebo "worked" in some way? Is that the kind of evidence you're talking about?Is that what you think? (Sorry but I don't think the passive tense is going to allow you to make this claim without evidence.)

If those in the placebo group show measurably better results than those in the control, then that would be evidence that the placebo has worked.

People have acupunture and it makes them feel better. People pray and it makes them feel better. People use magnets and power bands and they feel better. The evidence shows none of these things really work.......but they do.

I was asking for your evidence, not more of your opinions. But let me just ask something that should be easier: Where did you learn this?

Yeah, apparently there's not as much evidence for the placebo effect as is usually believed.

Same goes for acupuncture and prayer. Merely wishing for something to be true isn't sufficient to make it work and many times it isn't just the person who has the belief who's affected. Parents who think prayer or homeopathic medicines are effective ways to treat disease try to cure their kids that way, with disastrous results.

That's not what that article says - it says that around 2000 or so, some analysts cast doubt on a lot of previous studies that had shown a placebo effect because they thought that they weren't designed carefully enough (they quote a couple which didn't even have a control group) - the samples were small and they'd not ruled out some other possibilities. (Although it seems a bit misleading to me to concentrate on the ones without even a control group, because a lot of those difficulties are dealt with by the "many researchers" who did use a third control group).

Anyway, the analysis didn't say that there wasn't a placebo effect. It said that more careful studies were needed to demonstrate it. And then the article says that several more careful studies have since been carried out, and demonstrated that a placebo effect exists.

But if the truth is available, some people are going to find out about it, and that will presumably destroy the placebo effect. What constitutes forcing it on people? How do you even force someone to believe something? (I grew up in a heavily evangelical community; if it were possible to force me to believe in God, I'm sure they would have done it.) If I went to my doctor and asked for antidepressants and she told me she'd write me a prescription for Prozac if I really wanted it, but that in her studied opinion they don't work, is that forcing the truth on me or just being a responsible physician? (Again, assuming for the sake of argument that the placebo effect is the only reason Prozac works, which hasn't been established definitively.)

ETA: Personally, if my physician wrote me a script for a drug she knew was ineffective without telling me, I'd be furious.

Even more so with a drug that has very serious negative side effects, such as SSRIs. I mean, if it were OK to prescribe a medicine simply for a placebo effect, then responsible doctors would have to use something more like a true placebo and not a powerful drug that just doesn't work as claimed.

Here is another opinion about the so-called placebo effect:
http://www.sciencebasedmedicine.org/...lacebo-effect/

Of course a placebo effect exists - without it, homeopathy and a good bit of herbal medicine* would not be around.

Some things are subjective, to a degree. My wife thinks her homeopathic pills help with motion sickness. She thinks she gets less nauseous when she takes them. Nausea, like pain, if a hard thing to quantify. So, she thinks she is less nauseous, and can enjoy travelling more. Of course, it is all in her perception and the homeopathy does nothing physical. But - she feels better.

Granted, there is no placebo effect for actual physical effects, other than the fact that there is a connection between one's mental state and immune response, as mentioned in ganzfeld's link:

So, deal with the stress, the immune response might perk up a bit - it's not a huge response, but it exists. A placebo that makes a person feel less stressed can have that effect - even if the effect is minimal and due only to the person's idea that the "medicine" will work.


ETA: Obviously, placebos only work if the person believes and expects them to work. That's kind of the point, really. They don't work on everyone, and only work on things that are subjective (like pain levels), or that are tied to emotional/physiological connections. The are less "effective" on the latter.

Double ETA: I don't really get the logic of the article that Ganz lined to. He seems to be arguing that there is no biological placebo effect. Does anybody really argue otherwise? The impact is emotional - any biological effects are only a response to the emotional impact of the placebo.

I don't think that one sentence is a fair representation of the post at all.

This is very frustrating for me because we have about 22 factors that would not by anyone be considered "working" in any sense of the word (reporting bias, etc.). So let's call them A, B, C... V. Then we have about four things that people would consider working (mind over body, hormone action, etc.): W to Z. So scientists first have to eliminate A to V before they can even say that any of W to Z are happening. In the very very few cases in which A to V are eliminated or carefully accounted for, there's still a tiny effect. But at the same time the people doing these studies have already eliminated W, X, and Y. Z is still equivocal. Now comes the frustrating part. In conversations like this we have so-called skeptics claiming that W, X, Y, and Z have something to do with the placebo effect. No! Most of them have been ruled out. Even more frustrating is that we get another round of experiments that don't account for A to V completely and still claim their results are in favor of the so-called placebo effect. It's so stupid. No evidence from any of these experiments has ever shown that there is any benefit at all to giving someone a fake treatment, exclamation mark! Except that you can probably charge them something for it and keep them coming to your acupuncture clinic based on the incorrect and disproven claim that it "works" through the placebo effect.

Dr Novella's post is just a brief introduction to some of the problems with the claim that "placebos work" (which is not at all the same as asking whether there such a thing as a "placebo effect" exists).

Is decreased subjective perception of pain not a benefit?

Reporting less pain does not mean feeling less pain.

Ah. So you don't believe that the subjective experience is different, you think people are lying?

Yes, that is what scientists believe. Whenever someone can't describe the reality of a complex phenomenon exactly as it exists in nature, they must be lying.

I don't understand the eyerolling. I'm just asking questions. If the answer is no, you don't think they're lying, that's fine. I'm interested in understanding your view.

However, even if they don't feel less pain, but only think they feel less pain, then they feel "better" than they think they otherwise would. They feel relieved - this is only an emotional response, but it is still real, in a way. To me, that has always been my understanding of the placebo effect.

Not true. The placebo effect differs from spontaneous remission; in the former case, you feel better because you believe the treatment is working, while in the latter case you just happen to get better, regardless of whether you are aware you're taking a sugar pill. Even if the placebo effect didn't exist, spontaneous remission could account for a substantial number of cases in which people take some kind of quack medicine, recover, and believe the quack medicine cured them.

What Ganzfield is talking about (I think) isn't whether or not people think they're feeling less pain, it's that they think they should be feeling less pain and therefore report that they're feeling less pain regardless of whether or not they're actually feeling any relief- like how people with chronic illnesses that go to faith healers will announce that they've been healed even though they haven't because they think they should be healed.

No, that is an erroneous assumption. People do not, in fact, report what they think accurately in many situations, especially this kind of situation in which they have some feeling about how they think they are expected to respond. (Summing up 60 years of cognitive science: People don't always know what they think.)Yes, that is one common understanding that has remained an open question. What we do know is that the more strictly we control for the common errors (such as the reporting error mentioned, which is only one of many) the smaller the nonspecific effects.

I use the word nonspecific because it's a lot like dealing with UFO sightings. Before we consider that a specific set might be extraterrestrial crafts, we have to first eliminate the specific known causes of error. But eliminating all of them still does not mean we've shown ETs are here! What we are left with is an unidentified object, in medical treatment and experiment known as a non-sepcific effect, not necessarily a placebo effect.

I honestly don't want to get snarky, erwins, but it does no one any good when we have non-sequiturs such as: Are you calling all those people who saw these UFOs liars? No, of course not. But I think we can all agree that such questions miss the mark by so far that they shouldn't really deserve any comment.

When you referred to what people report vs. what they feel, I thought you might be saying that they were mis-reporting based on pressure or expectations, even if they were in fact feeling the same--i.e., they were lying. I don't consider that a non-sequitur, but if you don't want to answer my questions that's your choice. I don't have any kind of hidden agenda--I was genuinely trying to understand your position.

As in the case of all reported observations, including UFOs and ghosts and many other things, lying is one of many possibilities but not a necessary hypothesis. For example, I don't think people are usually even aware of the reasons they might select one box higher or lower on a score of ten. Nevertheless, many things very unlikely to be related to pain have been shown to influence such responses so it is not unlikely that there are at least several not yet accounted for. Many researchers have suggested ditching the term "placebo effect" for the more accurate "nonspecific effects". It seems to me the only reason that hasn't happened is because of the unfounded assertion that placebos actually have an effect.

One of the latter studies mentioned in the article first linked didn't rely on verbal reports of pain - it relied on requests for further pain relief from the three groups. (The only difference in that experiment was in how the saline drip that they were given was described to each of the three groups. All had the same physical treatment).

That's a "report" in a way, but it seems to me to eliminate most of any "I will say it hurts less because I think I should say that" factor. The patients were allowed to ask for extra pain relief whichever group they were in. Somebody could tell the doctor it hurt much less, and then ask for more pain relief than another patient who wasn't telling the doctor it hurt less. The article doesn't say, but it would be even better if the additional pain relief was measured by the number of button presses to administer an automatic dose, or some other means of eliminating the need to directly ask.

Not at all. And that's not the only factor anyway. If I think I've already received treatment is it even fair for me to ask for more treatment? All we know from these studies is that people who think they are being treated respond differently than people who do not. That's really nothing to base any treatment on. The bottom line is: If you have an ailment, get some attention. It doesn't require tricking the patient into thinking they're being treated (or, in the case of actual placebo-based studies where they do not actually lie to patients, have a 50-50 chance of being treated).

Also, it's such a gross oversimplification to say that this is the "I think I should say that" factor. What we're talking about is a barely statistically significant more likely to tick one box higher on a set of boxes. The reasons for that are complex and numerous. So to go back to the UFO example, that's like saying "well we eliminated the 'Venus on the horizon' factor so that's that". There are thousands of other possible lights besides Venus.

If there wasn't a placebo effect, it wouldn't need to be controlled for - people would control against groups that weren't receiving treatment at all.

The problem is one of definition, here, I think, and the level that you're looking at. You're taking the "placebo effect" and breaking it down into components, and then saying that because you can do that, the "placebo effect" doesn't exist.

There are times when a reductionist approach does remove the need for an overarching concept - "God", for one, is made up of a lot of unrelated and disparate ideas, some of which are valid and some aren't, and personally I think that the concepts are sufficiently different that the term itself is meaningless and actively confusing. But at other times, it's quite valid to group phenomena together and give them a name, even if you can break these down into more detailed entities. An "atom", say. A "carbon atom" is a useful concept even if, in reality, it's a collection of other waves and particles and ideas, and an "ideal" carbon atom doesn't exist.

I'd say that even if the "placebo effect" is a combination of the other things you've mentioned, it's still an effect. It still needs to be controlled for in a medical situation, and rather than writing out all of those other possible effects in full, usually it's useful to treat it as a single effect so that you can evaluate the effectiveness of the drug you're looking at.

That's no reason not to dive into it further, just as the periodic table is no reason not to look into atomic structure. But equally, the existence of electrons, and covalent bonding, and radioactive isotopes, doesn't mean that there's no such thing as a "carbon atom".

A carbon atom is made up of numerous smaller pieces that are linked together in a specific fashion and exhibits specific properties. The pieces that are cumulatively referred to as the placebo effect aren't really related to each other at all, they just get combined together under the same heading as a convenient catchall.

Yes, it needs to be used for what it was always and only intended for: in experiments to control for non-specific effects on subjective reporting, not as a BS excuse for using treatments that don't work in the hopes that they will do something that they, unlike experimental protocols, have never been shown to do. I wouldn't care what anyone calls it if they didn't abuse it.