I am annoyed by would-be skeptics and rationalists constantly drawing too many conclusions from very simple data, and then treating anyone who points out that the other conclusions are not valid as if that person disagrees with the one actual valid conclusion.
For example, take this study which legitimately does show that acupuncture is no more effective than fake-acupuncture. The linked writer's commentary is mostly fair enough, but they keep blurring out around the word 'placebo'. This is a really common thing amongst would-be rationalists. That writer is actually better than most, because they acknowledge in the first paragraph "an elaborate placebo whose effects, such as they are, derive from nonspecifice mechanisms having nothing to do with meridians, qi, or..."
The big problem is, that's not what the word placebo means. Placebo means something which does not have a physiological effect, not something whose physiological mechanism differs from that which has been proposed. And though the writer, in the first paragraph, specifically acknowledged that they were talking about a placebo which may have effects, as the article goes on they forget that because of using the slippery word placebo, especially when they slide on to describing it as "the placebo effect".
In the end, the writer, and many commenters, appear to have jumped from what the study actually shows - that acupuncture's results are identical to the results of very similar treatment that is not acupuncture but still involves lying down for half an hour having someone pay attention to you and touch you - to the conclusions absolutely not shown by this comparison, that "acupuncture does not work" and "it's all in the recipient's head". If the study compared against an actual placebo that would be a valid conclusion, but until it's shown that relaxing for half an hour has no physiological effects at all (not likely!) the study doesn't conclusively show "the placebo effect" at all. For that matter, it doesn't even show "acupuncture is no better than fake acupuncture", it shows only "acupuncture is equivalent to fake acupuncture as a treatment for chronic back pain".
Somehow, when I point this out, people think I'm arguing "acupuncture is scientifically valid," or "the study doesn't say acupuncture is not scientifically valid." I am saying neither of those things, what I am saying is that the study doesn't show "acupuncture is worthless", it only shows "fake acupuncture is just as valuable as acupuncture".
For the study to show that it's "the placebo effect" as people keep insisting, rather than "nonspecific other mechanisms that might apply to the treatment" it would require a study which tells some of the subjects that they're only going to be having fake acupuncture. You can't actually test acupuncture against a placebo because we can't magically skip someone forward half an hour in time and write "you just had acupuncture" in their brain, which would be the only true placebo for this purpose.
The phrase "no more effective than placebo" conjures the idea of sugar pills, which would be a totally different comparison. On that note, I like this analogy; comparing acupuncture to fake acupuncture and saying it's no more effective than placebo is akin to comparing bread pills against sugar pills in a starvation-prevention study, and concluding that bread is no more effective than placebo.
My favourite thing that came up on a forum where I was arguing this point, was someone who simultaneously pointed to two studies, one which showed "some placebos are more effective than others" and the other showed "placebos are no better than no treatment". So, A>B, A=C, B=C. Sound logic! [21:13] [0 comments]
|
|