Photo by syvwlch
Unless you majored in math or finance, odds are your formal education in statistics started and ended with the definitions of mean, median and mode somewhere in junior high. So it shouldn’t be surprising that most people are really bad at evaluating statistics from studies written for academic journals.
That includes the people who write the stories reporting those studies to the rest of us.
Relative vs absolute risk
If a study reports that increasing your consumption of wibbleberries is associated with a 33% decrease in your risk of a heart attack, that sounds like a big deal, right? Time to look for some wibbleberry pie recipes so I can live forever!
Except … what if the study was comparing a 0.000001% chance of having a heart attack to a 0.00000066% chance? Suddenly that sounds pretty small. Still kind of hard to get your head around though.
Try this: If you buy a lottery ticket you have one chance in 175 million of winning. Buy two tickets and you’ve got one chance in 87.5 million. Your relative chance of winning doubled, but your absolute odds went from 0.00000057% to 0.0000011%.
Whenever you see a story about an increase (or decrease) in the risk of dying, you have to check if they’re reporting a relative number or an absolute number. If it’s relative, remember that 33% more than a very small number is still a very small number.