UK Statistics Authority & the DWP - a clarification
With fresh reports about how statistics have been used by the Department for Work and Pensions this week, passing reference has been made to an investigation into DWP's statistics by the the UK Statistics Authority.
The impression has been given of a wide-ranging investigation into the accuracy of statistics released by DWP. This is not the nature of the concerns raised with the UKSA by Full Fact.
Given that it was us who first raised the issues being investigated, we thought it important to set out the specifics on the situation before any misconceptions about the situation spread too far.
Indeed, already it has been referenced in the press release sent out by the British Property Federation, in their notes to editors section as a statement of fact, despite us receiving no correspondence from them prior to the inclusion of this information in the press release.
Likewise readers of the Left Foot Forward blog will be well aware of the involvement of the Statistics Authority.
To clarify, our concern related to two specific press releases sent out by DWP in September regarding worklessness statistics, and that these statistics were sent out to selected members of the press, but not published anywhere on the DWP website.
Given that worklessness statistics are published as official national statistics, we were concerned that this may have breached official guidance as our blog at the time explained.
Therefore we wrote to the Chairman UK Statistics Authority Sir Michael Scholar to raise this point, and to avoid any confusion about the matter, our letter is published in full below, as is the UKSA response.
Ten days ago the UK Statistics Authority responded, telling us that at Sir Michael's request the matter was being investigated with DWP by the National Statistician.
So far as we are aware this investigation only relates to the two press releases in question rather than any other wider concerns about statistical releases by the Department, and we have as yet had no indication from the UKSA as to the outcome of the investigation.
"Dear Sir Michael,
I am writing to request your opinion on whether there has been a breach of guidance on official statistics by the Department for Work and Pensions in recent press releases.
My query is on behalf of Full Fact, an independent, non-profit organisation campaigning for greater transparency and accountability in the public debate.
On Monday 27 September several media outlets, including the Sun and the Daily Mail, reported figures from the Department suggesting that one in five claimants of three types of benefit (Jobseekers Allowance, Incapacity Benefit and Income Support) had received support for over five years.
However there was no mention of these figures on the Departmental website, nor was the press release published.
Though we have since obtained a copy of the press release which circulated the figures, the DWP media team explained that the data, while based on the Work and Pensions Longitudinal Study, was produced by fresh analysis of the database, and thus the statistics specific to the story were not publicly available when released to the press.
This is not the only time that such an approach has been taken to releasing statistics. On 13 September, a press release was circulated highlighting figures suggesting that one in 30 people of working age had never worked.
Again, neither the press release nor the figures were reproduced on the DWP website, and when my organisation requested the figures, we were simply sent the original press release because the figures were not publicly available.
Given that the 2009 guidance issued to Government departments by Sir Gus O'Donnell included a reminder that officials "must not selectively quote favourable data from any unpublished dataset", I wish to know if the press releases could be deemed contradictory to such guidance.
I note that you yourself have raised the issue of quoting unpublished data with the current Government in your letter to the Prime Minister dated 12 May, which stated:
"The Cabinet Secretary's existing guidance on statistical practice requires all officials and advisors, not only government statisticians, to observe the Code of Practice for Official Statistics; to consult professional statisticians when preparing publications; not to quote selectively from unpublished datasets".
In light of these comments I have enclosed a copy of the releases sent to the press in order to establish:
i) If the figures could be deemed official statistics, and if not, whether this was made sufficiently clear in the press release; and
ii) If the inaccessibility of the data itself is considered a breach of the guidance mentioned above.
Given the centrality of reforms to the welfare state in the debate on reducing the deficit, I am in no doubt that you will agree it is vital the public have confidence in the statistics that are being circulated on the issue.
Yours Sincerely
Patrick Casey"
The response we received on 9th November:
Dear Mr Casey
I understand that you have been chasing a reply to your letter dated 1
October to Sir Michael Scholar, Chair of the UK Statistics Authority and
received in our Newport office on 5 October. At Sir Michael's request, the
issues that you raise are being investigated with the DWP by the National
Statistician. I apologise for the delay and wish to reassure that we will
make a reply to you as soon as those investigations are completed.
Yours sincerely