The Skeptic's Health Journal Club: Clinical Significance versus Statistical Significance
In my last post we looked at some of the caveats one might consider when a study is shown to be statistically significant. I also made-up a really silly and stupid example example of a researcher looking at all kinds of variables until finally correlating as statistically significant, toilet seat position with future criminal behavior. Well, it didn’t take long at all for the real world media to present an actual example far sillier and stupider than my own.
See the recent article in the British Medical Journal which, after what might appear to be an excellent example of a statistical fishing expedition, pronounces thigh circumference to be predictive of heart disease. Far worse, though much funnier is the way the respectable organs of mainstream media picked up the article and ran with it, giving repeated breathless announcements throughout the news world such as:
Can Thunder Thighs Help Heart Health: ABC News
Large thighs 'may protect heart' BBC News
Fat thighs, bum may help you live longer: Global News
A big bottom 'is good for the heart': :NHS Choices
Why those fat thighs may help you live longer: Reuters
This being only a small sample of the dozens of similar news releases devoted to this issue.
I admit I was unable to imagine an example this stupid, though this real world “research” is presented as science. It is as though the media was telling me to not even try my hand as an amateur even at fantastical moronic pronouncements, they are the pros. An old Monty Python skit, a documentary on village idiots keeps popping to mind, especially the line where John Cleese looks thoughtfully into the camera and notes “Well a real blithering idiot these days can make anywhere up to 10,000 pounds a year”
