Frequent dancing causes a 76% reduction in the risk of dementia.
No, actually, it doesn’t. This is an observational study of two variables, which means it can’t tell which direction the causality goes, or whether there’s a common cause. And it’s quite obviously a common cause: poor health makes people disinclined to dance.
It’s possible that there’s still some beneficial effects from dancing, but the study does not show them, and the effect sizes won’t be anything near 76%.
This is one of the things I find astounding around here.
People here are so quick to poo poo the evidence afforded by anything but double blinded placebo controlled studies, and then turn right around and make claims with no backing at all. And a bunch of people mod them up for it.
You know that “it’s quite obviously a common cause”. And you know that the effect can’t be “anywhere near 76%”? On the basis of what do you claim to know these things?
People here are so quick to poo poo the evidence afforded by anything but double blinded placebo controlled studies, and then turn right around and make claims with no backing at all. And a bunch of people mod them up for it.
This is perhaps the wrong time to voice this complaint—because rejecting the claim made by the dance instructor is the obviously correct response. The study just doesn’t say that frequent dancing causes a 76% reduction in the risk of dementia. Even the people who wrote the study would cringe if they saw their work being used to make that claim! This isn’t “poo poo”, it’s “WTF?”!
Before I defend that statement—are you arguing that poor health does not cause dementia, that poor health does not disincline one to dance, that the types of poor health which cause these two things are different, or that the effect size somewhere is small enough that most of the observed correlation must be explained by something else?
(The claim of causation, by the way, was not in the study; it was added by Richard Powers.)
Change the phrasing here to be more charitable and I would say instead:
“There is an obvious common factor that explains some of the variance: dementia would stop people from dancing; the beginning stages would likely do so as well. Thus we should expect to see a correlation between dancing and good health even in the absence of dancing being preventative. Because of this expectation, seeing a strong link may be enough to locate the hypothesis, but it is likely to be some combination of statistical outlier and reversed causality. Thus if an effect exists, it is likely to be much smaller than the first measurement.”
And you know that the effect can’t be “anywhere near 76%”? On the basis of what do you claim to know these things?
I was ready to argue with you on this, because jimrandomh seemed to be applying a sensible heuristic here, namely: discount exaggerated reports of a study that say you can massively cut your risk of a chronic disease by doing an arbitrary thing. Normally there are only a few lifestyle factors that have such big effects: smoking, drinking, exercise, and diet. Maybe stress.
But of course frequent dancing is regular exercise. So, on reflection, the true causal effect could be on the order of 76% (though probably less). However, I predict it won’t be substantially more than the true causal effect of regular exercise in general.
I don’t agree with your general point; just because a claim isn’t explicitly justified doesn’t mean it has “no backing at all”. I would happily pooh-pooh double-blinded, placebo-controlled studies concluding that acupuncture can cure AIDS, even without “backing” my pooh-poohing, because it just isn’t plausible that acupuncture can cure AIDS and most people should have that as background knowledge (or at least be able to deduce it).
Still, we’re hardly talking about acupuncture & AIDS, so in this case your questions are fair. Dancing is exercise, exercise can have big preventive effects on chronic disease, and dementia is a chronic disease. Maybe dancing can reduce dementia risk a lot. I think an anti-halo effect from the sloppy argument in NancyLebovitz’s link made people underrate the hypothesis’ prior probability.
But of course frequent dancing is regular exercise. So, on reflection, the true causal effect could be on the order of 76% (though probably less). However, I predict it won’t be substantially more than the true causal effect of regular exercise in general.
From the study: “The physical-activity score was not significantly associated with dementia, either when analyzed as a continuous variable or when the study cohort was divided into thirds according to this score ”
That’s what I get for running my mouth without looking at the paper. Skimming it now, the “dancing [was] associated with a reduced risk of dementia” result smells like multiple testing throwing up spurious crap. There’s a dementia hazard ratio listed for 9 physical activities, and the ratio for dancing was the only one that was “significant” — but it’s only barely significant as the confidence interval goes from 0.06 to 0.99! And the hazard ratios for the other physical activities are all over the place, half of them less than 1.00, half of them at least 1.00. There doesn’t seem to be any real evidence of an association between exercise & dementia, taking this paper as a whole.
Actually, I now strongly suspect that many (though certainly not all) of the claimed benefits of exercise originated from the same error as seen here. But that is a claim that would require substantial research, to separate the real benefits from the fake ones and get the true effect sizes.
No, actually, it doesn’t. This is an observational study of two variables, which means it can’t tell which direction the causality goes, or whether there’s a common cause. And it’s quite obviously a common cause: poor health makes people disinclined to dance.
It’s possible that there’s still some beneficial effects from dancing, but the study does not show them, and the effect sizes won’t be anything near 76%.
This is one of the things I find astounding around here.
People here are so quick to poo poo the evidence afforded by anything but double blinded placebo controlled studies, and then turn right around and make claims with no backing at all. And a bunch of people mod them up for it.
You know that “it’s quite obviously a common cause”. And you know that the effect can’t be “anywhere near 76%”? On the basis of what do you claim to know these things?
This is perhaps the wrong time to voice this complaint—because rejecting the claim made by the dance instructor is the obviously correct response. The study just doesn’t say that frequent dancing causes a 76% reduction in the risk of dementia. Even the people who wrote the study would cringe if they saw their work being used to make that claim! This isn’t “poo poo”, it’s “WTF?”!
Before I defend that statement—are you arguing that poor health does not cause dementia, that poor health does not disincline one to dance, that the types of poor health which cause these two things are different, or that the effect size somewhere is small enough that most of the observed correlation must be explained by something else?
(The claim of causation, by the way, was not in the study; it was added by Richard Powers.)
Change the phrasing here to be more charitable and I would say instead:
“There is an obvious common factor that explains some of the variance: dementia would stop people from dancing; the beginning stages would likely do so as well. Thus we should expect to see a correlation between dancing and good health even in the absence of dancing being preventative. Because of this expectation, seeing a strong link may be enough to locate the hypothesis, but it is likely to be some combination of statistical outlier and reversed causality. Thus if an effect exists, it is likely to be much smaller than the first measurement.”
I was ready to argue with you on this, because jimrandomh seemed to be applying a sensible heuristic here, namely: discount exaggerated reports of a study that say you can massively cut your risk of a chronic disease by doing an arbitrary thing. Normally there are only a few lifestyle factors that have such big effects: smoking, drinking, exercise, and diet. Maybe stress.
But of course frequent dancing is regular exercise. So, on reflection, the true causal effect could be on the order of 76% (though probably less). However, I predict it won’t be substantially more than the true causal effect of regular exercise in general.
I don’t agree with your general point; just because a claim isn’t explicitly justified doesn’t mean it has “no backing at all”. I would happily pooh-pooh double-blinded, placebo-controlled studies concluding that acupuncture can cure AIDS, even without “backing” my pooh-poohing, because it just isn’t plausible that acupuncture can cure AIDS and most people should have that as background knowledge (or at least be able to deduce it).
Still, we’re hardly talking about acupuncture & AIDS, so in this case your questions are fair. Dancing is exercise, exercise can have big preventive effects on chronic disease, and dementia is a chronic disease. Maybe dancing can reduce dementia risk a lot. I think an anti-halo effect from the sloppy argument in NancyLebovitz’s link made people underrate the hypothesis’ prior probability.
From the study: “The physical-activity score was not significantly associated with dementia, either when analyzed as a continuous variable or when the study cohort was divided into thirds according to this score ”
That’s what I get for running my mouth without looking at the paper. Skimming it now, the “dancing [was] associated with a reduced risk of dementia” result smells like multiple testing throwing up spurious crap. There’s a dementia hazard ratio listed for 9 physical activities, and the ratio for dancing was the only one that was “significant” — but it’s only barely significant as the confidence interval goes from 0.06 to 0.99! And the hazard ratios for the other physical activities are all over the place, half of them less than 1.00, half of them at least 1.00. There doesn’t seem to be any real evidence of an association between exercise & dementia, taking this paper as a whole.
Actually, I now strongly suspect that many (though certainly not all) of the claimed benefits of exercise originated from the same error as seen here. But that is a claim that would require substantial research, to separate the real benefits from the fake ones and get the true effect sizes.