I found this looking for my regular risk tweeting, because a 50% lower grade seems like it might be a relative risk decrease masking quite a small absolute decrease Turns out, I was wrong But so was the headline!
-
-
Show this thread
-
I dug around for the study, and came up with this - researchers gave 100 students Fitbits to monitor their sleep and asked about their grades over a semester Those who slept more, did better academicallypic.twitter.com/OPPuLljEuN
Show this thread -
-
Now, this is pretty meaningless in and of itself anyway. People sleep well or poorly for a vast number of reasons, many of which impact academic acheivement as well Chances are, it's not the sleep per se that's ruining people's gradespic.twitter.com/1bGb42UbmG
Show this thread -
But then an interesting thing happened. I was looking through the paper to calculate an absolute risk reduction, and I realized there wasn't one!
Show this thread -
You see, being clever scientists, the researchers had analyzed their two numerical variables - hours of sleep and marks as a % - using linear regression, and hadn't looked at risks at allpic.twitter.com/fgjmV3dTfN
Show this thread -
Nowhere in the study is there the slightest suggestion that "College students who sleep less than 7 hours a night get 50% lower grades, study finds" So where did it come from???
Show this thread -
Well, I started looking around for other pieces about the research, and stumbled on this It's a quote from the press release that MIT put out about the study, from an academic unrelated to the research itselfpic.twitter.com/aEFRm97El0
Show this thread -
Now, this is pretty fascinating, because it seems what he's done is read off the regression line from the graph and compared the average of the two people who got ~6.5 hours sleep with the average of the dozens who got ~7.5 hourspic.twitter.com/yKTUbW16hQ
Show this thread -
So what we're really saying here is that students who get the ~least~ sleep perform, on average, about 50% less well than those who get the ~most~ And based largely on only two very poor sleepers!
Show this thread -
Thing is, the authors didn't even say this, probably because it seems like an anomaly based on variance due to their small sample size (remember, 88 completers in the analysis)
Show this thread -
But someone unconnected with the study read it and gave a really lovely quote, and that's what's hit headlines and probably at least 20,000 reader's minds Gotta love science communication
Show this thread
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.