Do we have a "stats that make you cringe" thread?
In my current line of work, I'm frequently reading reports of studies that claim some behaviour or other "increases risk of death by 30%" or some other figure.
I mean, surely we're all at 100% risk of death sooner or later? There must be a better way of phrasing what these study results mean...
There is. Relative and absolute risk are the correct terms. An absolute risk is saying you have a 1 in 10,000 chance of being killed by lightning (in your entire life) or saying that 9 in 10 men will have evidence of prostate cancer by the age 60.
Relative risk is the difference between two groups, so you could say wearing a tin hat increases the likelihood of getting struck by lightning by 100%, so the absolute risk for tin hat wearers is now 2 in 10,000.
The percentages are worthless and strongly recommended against as they give no idea as the actual risk, saying eating cheese increases the risk of a particular cancer by 30% sounds horrific and you must stop immediately, but if the incidence of that cancer is 1 in a 1,000,000, then 30% is insignificant.