Percentages are useful but they’re also misleading. Saying there’s been a 100% rise in the incidence of heart disease amongst young women may sound shocking but if the incidence only used to be 1 in 500,000, the new figure of 2 in 500,000 it’s really not that bad, but then that’s not a news story…

The other, less well appreciated problem with percentages is the reporting of changes in percentage, here’s an email sent ’round my dept that does a good job of of explaining it (hope you don’t mind Gary if you’re reading this by some strange coincidence) …

If something - exam results, say - goes from 10% to 11% it is NOT “up 1%”.

I hope you can see, because the number are simple, that it is an increase of a tenth, which is 10% - or, in that example, “one percentage point”.

Exactly the same principle applies if you go from, say, 58.1% to 59.7%:
It is not “an increase of 1.6%” it is an increase of 1.6 percentage points.

You work out the % increase by dividing the difference by the lower figure - which in that case gives 2.75%.

So a rise from 20% of something to 40% is NOT “a rise of 20%” it’s a rise of 100% - in other words, a doubling.

Thing is, saying a change from 1% to 2% is a rise of 100%, whilst accurate, is pretty sensational. My preference wherever possible would be to use proportions i.e. ‘1 in however many’. For a change the best idea is proabbly to explicitly state how the proportion has changed eg. ‘up from 1 in 100 to 2 in 100′.

Obviously I’m assuming the goal here is to give people an accurate perception of the numbers not to have the most sensational headline.