35% vs 300%

Stats Chat reader Elizabeth Paton-Simpson emailed us about what would have been a worthy Stat of the Week nomination. It’s from back in June, so isn’t eligible, but I wanted to share it with you:

Elizabeth wrote the following to the Sunday Star Times:

An item in the Sunday Star Times on 2 June was headlined: “Anaesthetics triple risk of dementia for elderly”. Likewise, the opening paragraph proclaimed: “Having a general anaesthetic may triple the risk of dementia in elderly patients, a study suggests.”

Trained to be sceptical of media claims, I read on and found reference only to a 35 percent increase. Checking the purported source revealed that the Sunday Star Times had altered the summary in The Times from “may increase … by a third” to “may triple”. So please tell Grandma she need not be alarmed – it’s just another example of journalistic innumeracy.

The Sunday Star Times responded:

Yes, you’re correct that the Times later corrected their article however we took their wire service copy which originally said “triple”. I’m sorry for any alarm this may have caused! And, of course, you are right that this should have been picked up in our subbing process.

Unfortunately, the statistic remains uncorrected elsewhere, such as The Australian.