The UK's statistics watchdog has rebuked a minister over his handling of controversial figures on benefits claimed by immigrants.
Sir Michael Scholar, head of the UK Statistics Authority, has written to Work and Pensions Secretary Iain Duncan Smith, questioning the way he released the figures almost a week ago.
Mr Duncan Smith has already responded, defending how the figures were presented and saying he didn't break any rules.
The rebuke relates to a report published by the Department for Work and Pensions (DWP) last Friday which was the first ever stab by a government at calculating how many foreign nationals claim some benefits.
The report would inevitably trigger public debate - and the Daily Telegraph published a comment piece by ministers hours before the figures were released.
In the article Chris Grayling and Damian Green, ministers for employment and immigration respectively, referred to figures in the report. They said there were 371,000 people who were foreign nationals when they entered Britain who were today claiming benefits.
Referring to a portion of those, they added: "We've already identified some with serious question marks over both their right to benefits and their immigration status. Investigators are calling to see them.
"We're also working on urgent plans to streamline the rules so that we can stop benefits immediately. Under the regulations we inherited, it takes nearly three months in a case like this. That has to stop. All of this represents an almighty mess [inherited from Labour]."
The story was immediately picked up by other national newsrooms and it was briefly a lead item on the BBC. Chris Grayling came on to the News Channel to give his take on what it all meant.
The DWP figures were published as a "research report" rather than as a release of "official statistics". Does that matter? Yes it does, say Sir Michael.
Official stats are numbers that have been tested, and tested again, by professionals who work within strict rules. Their reports are presented independently of ministers, the opposition, policymakers or anyone else for that matter. If it's an official stat, then it's a fact.
In his letter, Sir Michael says the numbers should have been prepared as official statistics because of the obvious public interest.
And that basically means that ministers should not have had their say before the figures had been tested and published for all to see.
Sir Michael said: "Many users have treated them as official statistics, and have assumed that they should have been published in accordance with the Code of Practice, which would, amongst other things, have prevented government ministers from issuing a political commentary on the statistics ahead of their publication. The Statistics Authority shares that view.
"These statistics are both highly relevant to public policy and highly vulnerable to misinterpretation.
"There are some important caveats and weaknesses that need to be explained carefully and objectively to Parliament and the news media at the time of publication.
"This is, in our view, best done by official statisticians producing a statistical release in accordance with the Code of Practice."
So what were those weaknesses? Well one point that wasn't apparent the night before, but became clear once the full data was studied the following morning, was that foreign nationals may be less likely to claim benefits than British born workers.
Sir Michael Scholar now wants the DWP to agree that future figures on benefits and foreign workers should be handled as official statistics so that they can be presented in a factual, politically neutral way.
He also wants the raw data published so that experts outside government can do their own number-crunching.
In his response, Iain Duncan Smith has defended the quality of the statistics, noting that he has not been accused of a breach of the code of practice.
He said: "The release… was prepared by the department's statisticians to appropriate professional standards. The report includes all relevant information about the sources of the data and any caveats or limitations.
"I note your point that if the exercise were to be repeated then it would be appropriate for subsequent releases to be as official statistics, however, we have no plans at present to repeat this analysis."