I don't understand how people can think scientists don't know this. From the article:
People who reported they slept 6.5 to 7.4 hours had a lower mortality rate than those with shorter or longer sleep. The study, published in the Archives of General Psychiatry in 2002, controlled for 32 health factors, including medications.
Of course controlling for confounders does not magically mean their study is perfect - there are always confounders that are either unknown or unable to be measured, particularly in retrospective/observational studies like this. But it's not nothing.
I think it's somewhat hard to necessarily point to the direction of association pointing the way the paper asserts its pointing, even leaving
I would have been much more impressed by using sleep as a time-varying exposure, rather than a time-fixed exposure, but the study is a rather old one.
Lumping all reports from 9.5 to 16.5 hours makes me a little nervous however - that's a range from "sleeps more than average" to "sleeps the vast majority of the day, and enough that a full time job is impossible". I'd assert those are very different people.
>>Posting those 5 words doesn't dismiss an argument. While it doesn't equal causation, it's a sure as hell cue that's something funny might be going on.
It may be a cue, but cues themselves are not actionable. They simply mean that more research is needed until causation is found.
People who reported they slept 6.5 to 7.4 hours had a lower mortality rate than those with shorter or longer sleep. The study, published in the Archives of General Psychiatry in 2002, controlled for 32 health factors, including medications.
Of course controlling for confounders does not magically mean their study is perfect - there are always confounders that are either unknown or unable to be measured, particularly in retrospective/observational studies like this. But it's not nothing.