Last month FactCheck waded into a row over the government’s decision to cap benefits at £500 a week for most couples and single parents and £350 for single people.

The work and pensions secretary, Iain Duncan Smith, announced that the mere threat of the cap being brought in had already nudged some benefits claimants towards finding a job.

Mr Duncan Smith told the Daily Mail:

“Already we’ve seen 8,000 people who would have been affected by the cap move into jobs. This clearly demonstrates that the cap is having the desired impact.”

In fact, the government research quoted by Mr Duncan Smith made it clear that, while 8,000 more people did move into work, there was no way of knowing that this was anything to do with Jobcentre staff issuing warnings about the looming benefits cap.

Mr Duncan Smith’s own researchers said the figures were “not intended to show the additional numbers entering work as a direct result of the contact”.

Over any period of time, some unemployed people will move into work, regardless of whether a benefits cap is about to be implemented. Experts agreed that a fall in unemployment didn’t prove that government policy was working.

In a second piece of analysis, the researchers said fewer households would be affected by the cap than was initially estimated, but put this down to “normal benefit churn” rather than a change in claimants’ behaviour.

The minister had no excuse for making this slip, as FactCheck had already pulled him up on it.

Mr Duncan Smith first made this claim in July last year, saying:

“The benefit cap is already a success and is actively encouraging people back to work.”

We established that there was in fact no statistical evidence for this kind of optimism. A DWP spokesman simply said: “The secretary of state believes that the benefits cap is having an effect.”

But that didn’t stop Mr Duncan Smith repeating the same claim in the Mail nine months later.

Now Andrew Dilnot, chairman of the UK Statistics Authority, has backed our Fiction verdict.

In a series of open letters, Mr Dilnot said Mr Duncan Smith’s claim “is unsupported by the official statistics published by the Department on 15 April”, and criticised DWP for releasing the figures Mr Duncan Smith drew on in his comments to the Mail.

Mr Dilnot said: “The statistics do not comply fully with the principles of the Code of Practice, particularly in respect of accessibility to the sources of the data, information about the methodology and quality of the statistics, and the suggestion that the statistics were shared with the media in advance of their publication.”

This is not the first time Mr Duncan Smith’s department has received a rap on the knuckles over its use of statistics.

Mr Dilnot says that the department had assured him after an earlier complaint that “senior DWP officials had reiterated to their staff the seriousness of their obligations under the Code of Practice and that departmental procedures would be reviewed”.

He adds: “The Board of the Statistics Authority would welcome further assurance that the working arrangements within the department give sufficient weight to the professional role and public responsibilities of statisticians.”