Statistics User Forum

last person joined: 2 months ago 

The Statistics User Forum (SUF), supported by the Office for National Statistics, aims to be the strategic voice of users of official statistics. The Forum acts as an umbrella organisation for user groups, each of which covers a particular area across the spectrum of official statistics. Through the Forum we enable open and constructive dialogue between users and producers of official statistics and encourage communication across the user community.

Publishing Official Statistics on Polling Days 

14-11-2017 16:42

Dear SUF members,

Please see this letter from the National Statistician about the publication of official statistics on national polling days. I am uploading a separate file summarising the policy. I would welcome your views on any of this - by close of Thursday 7 December, please, so that I can reply John Pullinger by his 12 December deadline.

Thank you,

Paul Allin
Chair, SUF

Statistics
0 Favorited
8 Views
2 Files
0 Shares
210 Downloads

Related Entries and Links

No Related Resource entered.

Tags and Keywords

Attachment(s)
pdf file
Polling Day Policy   163K   1 version
Uploaded - 14-11-2017
For your comments, please
pdf file
Letter from National Statistician   396K   1 version
Uploaded - 14-11-2017
About publishing official statistics on polling days

Comments

18-12-2017 13:01

SUF members may be interested to see the following response made on this, on behalf of SUF and the RSS:

SUF and RSS response to ONS Consultation on the publication of official statistics on national polling days
We support Alternative D in the policy paper, for two reasons:
1) The need to avoid statisticians being perceived as taking selective decisions regarding which statistical releases will be of significant public interest on polling day
2) The perspective of users of official statistics.

From the perspective of users of official statistics:
• users value certainty in pre-announced release dates;
• where pre-announced release dates need to be changed, albeit for good reasons, then users value as much notice as possible (and reassurance that the date will not be further changed);
• users have many different interests across the range of official statistics: these are not well mapped by producers, but the diversity of usage puts into question the concept of “statistics of significant public interest”;
• users have a strong stake in preserving the integrity of the UK statistical system;
• one way in which integrity could be called into question is if there is a suggestion (even a misunderstanding) that release dates might be manipulated, or that the releases of some statistics are treated differently from others.

Additionally, we have considered the view that general election days take precedence over the release of official statistics. Although we do not have direct evidence for this so can offer only conjecture, it is clear to us that users can point to many examples where the release of a set of official statistics could chime with a topical election issue, within a constituency if not nationally (e.g. health statistics could be relevant where the planned closure of a hospital is at issue).

Another potential complication is that official statistics tend to be released at 09.30am, so that some votes could be cast before a set of statistics is released, while votes cast later in the day could be influenced by the statistics.

Under the Fixed-term Parliaments Act 2011, we have a due date for each forthcoming UK parliamentary general election, although this can be brought forward. So, assuming no early election is called, we currently know that the date of the next general election will be 5th May 2022.

Taking account of all these points suggests a policy as follows:
(a) no official statistics to be released on polling days (which in particular means that no statistics should be scheduled for release on 5th May 2022, unless that date is changed);
(b) as soon as polling days are known, all official statistics that have been pre-announced for release on that day should be re-scheduled to be released on the following day, and the list of release dates updated immediately.

This is closest to Alternative D in the policy discussion paper.

Submitted 14 December 2017