The Office for Statistics Regulation (OSR) has conducted a consultation on proposed changes to the Code of Practice for Statistics, on behalf of the UK Statistics Authority.
OSR is grateful for responses to our consultation on release times. You can view the results on our release times review publication.
Reason for change
During the coronavirus pandemic, we have been reminded of the importance of ensuring statistics are available to everyone at the same time. It is vital that they are shared in ways that reassure users about their integrity and independence.
The Director General for Regulation granted exemptions in the early months of the pandemic for some official statistics to be able to be published at times other than 9.30am, as required in practice T3.6, while remaining transparent and orderly. The Office for National Statistics (ONS) and the devolved administrations are publishing market sensitive statistics at 7am. Public Health Scotland (PHS) and other producers are publishing some COVID-19 related statistics later in the day at agreed times, as soon as prepared and quality assured.
In considering a change to the release practice, we spoke with heads of profession for statistics about their views on how to best serve the public good and ensure orderly release. Comments received from producers, including ONS and the devolved administrations, showed strong support for a standard release approach but with some flexibility allowed. The release time of 9.30am was considered broadly preferable for most releases, although there was interest in some latitude for the standard time being set differently accordingly to local needs. We received feedback from some producers that early release outside of usual working hours was a substantial concern for staff, plus a higher risk of release breaches. Those involved in the release of pandemic health data were keen to have the opportunity to release in the afternoon for rapid release statistics. A clear message was to remain orderly and transparent in the manner of release, to ensure and protect statistical independence.
We also received comments from users of statistics and other stakeholders. Of the 20 responses, three quarters of the stakeholders expressed a preference for release at 9.30am. A quarter were positive about early release (one found early release helpful but recognised the burdens imposed), while half cited negative impacts of 7am release.
While practical and personal challenges experienced with early release were reported, our overall view is that some flexibility in the timing of release has better served the public good, by enabling statisticians to speak for their statistics. Publishing market sensitive statistics early has had the benefit of preventing speculation about key economic statistics and enabled them to directly reach a far wider audience. Publishing the PHS COVID-19 weekly bulletins and ONS Coronavirus infection survey reports at noon has meant important statistics have been available at a consistent time with sufficient opportunity to assure their quality.
The current practice in the Code sets a single time of 9.30am. It is now clear that there are situations in which producers would sensibly want to depart from this time, and the Code should recognise the value of this alternative timing. In proposing a change, we seek to enable some flexibility, support producer judgement, and maintain orderly release.
We propose to change practice 3.6 of the Code of Practice. It currently states:
“Statistics should be released to all users at 9.30am on a weekday”
“Statistics should be available to all users at the same time. The standard release time is 9.30am on a weekday, but there may be circumstances when it better serves the public good for statistics to be released at an alternative, pre-announced time. Exceptions can therefore be granted by the Director General for Regulation. The time of release, whether it is the standard release time of 9.30am, or another time granted by the Director General for Regulation, should be announced in advance and adhered to consistently.”
As with the current practice, this option maintains 9.30am as the standard release time for statistics; but it differs in that it also enables alternative releases times to be granted where the Director General for Regulation agrees that it is in the public interest to do so. The alternative times would be applied consistently from the point of the exception being granted, unless it is agreed by the Director General that the release time be further changed.
“Statistics should be released in a way that is transparent, independent, and orderly, in a manner that best serves the public good and is agreed by the Chief Statistician/Head of Profession for Statistics. A consistent release time should be used by an organisation for all its statistics. Exceptions may be granted by the Director General for Regulation where it is in the public interest to publish at another time. The time of release, whether it is the standard release time, or another time agreed with the Director General for Regulation, should be announced in advance and adhered to consistently. The needs of users, the wider benefit to society and the wellbeing of staff should be factored in when deciding the time of release.”
Rather than setting the explicit time of release, this option establishes the core values that should be met in ensuring the orderly release of the statistics. Again, the importance of a consistent and transparent release time is underscored but this option allows for a producer body to decide the most appropriate time of release within its own context. This time would need to be applied consistently across all the producer’s statistics. Some further flexibility is provided through allowing exceptions to be granted by the Director General for Regulation, as in option 1.
The decision about the appropriate time of release may require balancing competing views and needs. We believe that this decision should be made in a way that protects statistical independence and guards against vested interests. The choice should ensure orderly release that has consistent timing and is in line with the legislation. It should preserve public confidence through transparency both before and at the time of release. It should enhance expert use, support the wider public accessibility of statistics, and enable wide, accurate media reporting of the statistics.
We further propose to change practice T3.1 to require that the release time, as well as the date, be pre-announced. In addition, we propose to remove the word ‘official’ after ‘regular and ad hoc’. The practice would become:
“The release approach of both regular and ad hoc statistics should be pre-announced through a 12-month release calendar, giving the specific release date and time at least four weeks in advance where practicable.”
This consultation opened for 12 weeks from 28 September 2021. We are now considering those responses and will report on next steps shortly.
If you have any queries, please contact OSR by emailing firstname.lastname@example.org
Please be aware that any information provided in a response to this consultation could be made publicly available if requested under a Freedom of Information request.