Small Data: A question of 0.9 percentage points
There was a slap on the wrist for HM Treasury last week for not being clear enough about which measure of inflation it was using, writes Anthony Reuben.
At the centre of the argument was a 0.9 percentage point difference about whether "frozen" regulated train fares next year would go up by 1.6% or 2.5%.
On a £2,000 annual season ticket, that's the difference between a rise of £32 and a rise of £50.
In a press release titled "Government extends freeze on rail fares to 2015", the Treasury announced: "No regulated rail fares will rise by more than inflation in 2015."
But it's not until seven paragraphs down that it mentions that the rate of inflation being used is the Retail Prices Index (RPI) and not the government's preferred measure of inflation, the Consumer Prices Index (CPI). An earlier version of the release just said "inflation" without distinguishing between the measures at all.
The rate of inflation used is the one for July the previous year. In July 2014, RPI was 2.5% while CPI was 1.6%.
In March 2013 the UK Statistics Authority (UKSA) decided RPI should no longer be called a National Statistic because of issues with the way in which it is calculated.
Nonetheless, RPI is used to determine how much regulated train fares can rise each year.
The UKSA sent a letter responding to a complaint from the Royal Statistical Society's RPI and CPI user group, saying it had been assured that: "Arrangements have been made to update the press release."
In the June 2010 Budget, the government announced that it was switching from uprating pensions and benefits from RPI to CPI, which the House of Commons Library estimated would save it £5.8bn this year.
The RPI almost always rises at a faster rate than the CPI. The RPI is also used for changes in rents on social housing, although that is moving to CPI next year.
The campaign group Rail Future reckons rail fares should go the same way.
Subscribe to the BBC News Magazine's email newsletter to get articles sent to your inbox.