Statistics and research: release practices protocol

This Protocol sets out how the Chief Statistician and, with their authority, other members of the Government Statistical Service in the Welsh Government, and associated bodies, will carry out their responsibilities in respect of release practices.

Authority for the Protocol derives from the Statistics & Registration Service Act, 2007 which received Royal Assent on 26 July 2007. It is written in accordance with the Code of Practice for Statistics, published by the UK Statistics Authority (The Authority), in February 2018. It also corresponds to the Office for Statistics Regulation (OSR) guidance ‘Publishing official statistics and National Statistics’, published in October 2018.

In building trust in official information, it is important that clear and principled release practices are integral to the identity of all official statistics. The emphasis of this Protocol is on transparency, key principles and compliance with a small number of widely relevant rules.

Accredited official statistics cover government data produced to the high professional standards set out in the Code of Practice, and are identified by the accredited official statistics logo. However the principles underlying this Protocol of open release practices should apply to all official statistics produced by the Welsh Government.

Welsh Ministers, including the minister responsible for statistics and in consultation with the Chief Statistician, have set out the rules governing pre-release access via the Pre-Release Access to Official Statistics (Wales) Order 2009.

The National Statistician as Head of the Government Statistical Service, and in consultation with the Chief Statistician in the Welsh Government, will have responsibility for the professional integrity of those outputs comprising official statistics. The National Statistician, in conjunction with the UK Statistics Authority, will also set professional standards for official statistics, including standards for release arrangements and quality assessment. The Authority has the role of determining whether or not a statistical output meets those standards via a formal accreditation process.

The Chief Statistician in the Welsh Government, in consultation with the National Statistician, will promote and ensure adherence to the Code of Practice for Statistics and this Protocol, and seek to resolve any issues concerning interpretation of the Code and Protocol in the Welsh Government. The Chief Statistician is the authoritative source of advice to ministers on their interpretation. Final responsibility for the content, format or timing of release of official statistics will rest with the Chief Statistician.

Those producing statistical outputs(a) must be protected from any political pressures that may influence the production or presentation of the statistics.

The Chief Statistician, as Head of Profession for statistics, has sole responsibility for deciding on the content and timing of statistical releases(a).

All statutory obligations and internally endorsed guidelines, which govern releases, will be followed.

All government statements issued alongside official statistics, which refer to or are based upon them, will:

• contain a prominent link to the statistical release and clearly refer to the source of the statistics
• be labelled clearly as policy statements or ministerial statements, and are readily distinguished from a statistical release
• meet the professional standards outlined by the Welsh Government

(a) References to statistical releases or outputs throughout this protocol refer to all formats of statistical information published by Welsh Government. It’s not just limited to statistical first releases including bulletins, StatsWales data, headline pages, interactive tools or websites etc.

Statistics will be presented impartially and objectively.

A fixed format for presentation, taking account of user needs, will be adopted for regularly recurring releases. This will be sufficiently flexible to allow explanation of aspects of the data as they vary from period to period.

Significant changes to the content of outputs must only be done on the basis of user needs, and after discussions with them, whilst ensuring that accessibility standards are still met. Reasons for the change must be publicly available and, where possible, pre-announced. Users will be consulted on any proposed changes to the standard publication arrangements.

Unplanned or selective release of parts of research is to be avoided, and where it occurs it must be for statistical rather than policy reasons.

Official statistics should be made equally available to all, at the same time, with limited exceptions. All statistics should be released in formats that best meets the needs of users of those statistics, including those with impairments, and which promote widespread access and informed debate. 

Statistics should be made available in as much detail as is reliable and practicable, subject to legal and confidential constraints, offering choice and flexibility in the format according to the level of detail required by the user.

Statistics should be publicised in ways that enable users to identify and access information relevant to their needs.

Official statistics must be disseminated in forms that enable and encourage analysis and re-use. Data published should be in-line with the Welsh Government’s open data plan and use Regional geography codes.

All information relating to the release of official statistics must be publically available.

A programme of statistical outputs will be maintained and regularly updated for a full year ahead in line with the Code of Practice for Statistics.

• Statistical first releases are used to publish new data.
• Statistical bulletins provide more detailed secondary analysis of datasets.
• Headline statistics are used to provide short standalone analysis or signpost to analysis produced by another department.
• Data releases are new statistics which are made available via StatsWales or an interactive tool (such as Microsoft PowerBI or Tableau) only. These are also announced in-line with the Code of Practice for Statistics.

Occasionally statistical articles may be produced which provide an analysis of particular methodological issues or experimental analysis. These are generally reactive and will be included in the upcoming calendar as soon as any plans are developed to produce an article.

Interactive websites are also produced to support some statistical releases. These websites could also represent the first publication of a dataset.

Statistical outputs will be translated in accordance with the Welsh Government’s requirements under the Welsh Language Standards.

Outputs will be in line with government digital publishing guidance, including the standards for accessibility.

Early access to official statistics can only take place where practices to prevent any opportunity (or perception of opportunity) for personal advantage have been adopted. All such pre-release arrangements will be made publicly known in accordance with the Pre-Release Access to Official Statistics (Wales) Order, 2009.

Individuals given early access will not seek changes in publication dates or to the content or format of the material they receive.

Attempting to profit from early access to market sensitive information (e.g. insider trading) may fall within the civil regime of market abuse as well as amounting to a criminal offence under UK law.

Where privileged early access to market sensitive statistics is provided, it will occur not more than 24 hours before release. In normal circumstances, pre-release access to nonmarket sensitive information will be supplied to those entitled to pre-release access no longer than 5 working days before publication.

Notes

This period applies to regular working weekdays and will be extended accordingly if such a period spans a weekend and/or a public holiday.

Only the Chief Statistician in the Welsh Government may authorise an extension to this period, if it is justified by special circumstances. Operational delays will not normally be a reason to extend the period for privileged access.

Currently no official statistics produced by the Welsh Government are deemed as market sensitive.

Privileged access for Welsh Ministers

Ministers may have early access to statistical outputs, along with their officials. The sole purpose is for ministers to be able to respond completely when questions arise at the time of publication.

Privileged access will be strictly limited and any such access will be limited to the minimum amount of time required to prepare a useful response to the data when it is released. In normal circumstances, this will be no longer than 5 working days before publication. This Protocol works in accordance with the Pre-Release Access Order (Wales), 2009, Schedule 1.1 to 1.3, in setting out who is normally entitled to have privileged early access and for how long.

Security markings and a simple reminder of the legitimate use of official statistics under privileged access should accompany all material supplied before release.

Emerging findings

Where official statistics are derived from an extensive, ongoing programme of work, for instance on an annual production cycle, ministers or others who are not part of that particular statistical production process may need to be informed of emerging findings well in advance of the release, at a stage when the final statistics are not available, in order to consider policy implications. In such situations the Chief Statistician in the Welsh Government, in consultation with the National Statistician if necessary, will assess whether exceptional early public release of these emerging findings is required.

This does not apply to statistics with periodicity of less than a year; these releases are produced in such short timescales that the concept of emerging findings cannot apply. Preliminary findings to these regular statistical outputs will not be released.

Exceptional events

Exceptional arrangements can be made for key economic events. For example when the date of the Budget or pre-Budget report is announced each year the National Statistician will determine whether, for the sole purposes of preparation of the Budget statement, there is a compelling case on operational grounds for exceptional early access to certain data for a limited number of Treasury officials. This might include official statistics produced by the Welsh Government. This will usually be provided, and will be retrospectively noted on the affected statistical releases.

The same procedures will apply for other key events which the National Statistician deems to require exceptional treatment.

Quality assurance

Early access to unreleased official statistics can be given to people able to offer useful ideas, comments and criticism on the quality and presentation of the information, or on any significant related issues. Where this is done, the team producing the statistics should keep an accurate audit trail of the distribution, including what is sent, to whom, when and where. Those receiving unreleased material should be advised of their responsibility in respect of confidentiality, and warned of possible legal penalties that might apply to wrongful release.

Where possible such access will be granted in enclosed environments where no material is physically or electronically taken away by those having sight of unreleased material. Use of electronic file sharing systems where access can be limited is encouraged.

International obligations

International treaties and agreements oblige the UK to provide certain data to international statistical organisations for the purposes of compilation of international comparisons and outputs. This may involve the early release of statistics relating to Wales in order to compile UK-wide statistics.

Any requests for privileged pre-release access to official statistics by international statistical organisations will be considered case by case by the National Statistician and the Chief Statistician in the Welsh Government.

Management information

Where official statistics are derived from sources primarily used to manage and inform government processes, a number of individuals will have access to the data sources in the course of their normal work. For practical purposes, this amounts to a form of early access and needs to be recognised as such and controlled accordingly.

Use of administrative or management information that is used for the production of official statistics should follow the National Statistician’s guidance on management information and official statistics.

In terms of release practices, the following two points should be noted:

• to provide confidence that wide access within government to this information before its public release does not affect the form or context around which release of the resulting official statistics occurs, publication dates will be scheduled well in advance and publication formats will be rigidly adhered to
• those who have access to these data should avoid any public comment that would damage the integrity of that official statistics release

Release management

If there is an operational need for early access for organisations sharing responsibility for the distribution of statistical releases, early access will be granted where the Chief Statistician in the Welsh Government is satisfied that suitable practices have been adopted to limit opportunity for personal gain, and where arrangements are sufficiently transparent.

Embargoes

As a general rule, embargoed access is not provided to the media.

Under exceptional circumstances the Chief Statistician may determine that embargoed access is appropriate or necessary. In such cases embargoed access may be given to accredited journalists, where it is seen as necessary to provide them with a period of time to assimilate and comprehend the data, in order to provide for informed comment at the time of release, or if it is required to distribute material to the media in time for publication.

Embargoed access cannot apply to market sensitive releases. It can only apply to the release of non-market sensitive data where there is limited opportunity for personal gain by those who have early access. It will only be given for:

• complex reports, the value of which will be better understood by the public if accompanied by informed and considered comment at the time of release
• compendia or value-added reports which are not putting new basic data into the public domain

Accidental and wrongful release

Accidental or wrongful release by those with early access will result in action to prevent recurrence.

Any accidental or wrongful release will be reported to the Chief Statistician immediately after it is noted, so that appropriate action to limit loss of confidence is taken quickly.

Accidental or wrongful release includes providing any indication of the content of the release, including suggestions of the size or direction of any change or a substantive description, as in ‘favourable’ or ‘unfavourable’.

Where there is clear proof that wrongful or accidental early release of an official statistic has occurred, the Chief Statistician in the Welsh Government may judge it necessary to arrange for immediate release of the statistic. Release to selected national media will be defined as public access for this purpose. Notification will also be provided on the website, via social media, and by direct communication to users.

Where wrongful or accident early release of an official statistics has occurred the Chief Statistician in the Welsh Government will report this to the UK Statistics Authority as soon as is practicable and compile a formal breach report.

Publication timetable

For any sensitive statistics and for frequent regular outputs where the production timetable can be predicted far in advance, the exact day of publication will be provided at least six months in advance.

For all other outputs the month will be provided, either when the data have been collected or six months in advance, whichever is the earlier. Exact publication dates will be provided as soon as possible, but no less than four weeks before release and without reference to the specific information to be released.

Under exceptional circumstances the Chief Statistician may agree for those outputs described in para ii to be pre-announced less than four weeks in advance.

Tables of data on StatsWales which do not represent the sole or initial method of release of data will not in general be pre-announced. These would generally include more detailed breakdowns of data or tables to support statistical releases which cannot be produced according to the primary production timetable.

Any alteration to a pre-announced date will be made public as early as possible and accompanied by an explanation for the change. This will include notification on the website, via social media (for key outputs or where changes are made at short notice), and by direct communication to users.

A programme of statistical outputs will be maintained and regularly updated for a full year ahead.

Responses to ad-hoc statistical requests that have been answered will be published in batches on a fortnightly basis.

Releasing statistics

In general, with a small number of exceptions where they rely on data published by another source, all releases will take place at 9:30am on a weekday.

The method of release will be consistent, including availability of the release on the internet and notification via social media.

Users will be advised about the release arrangements well in advance.

Any errors discovered in statistical reports will be corrected promptly, stakeholders will also be notified. This will be in accordance with our revisions policy.

Welsh Government statisticians, who are responsible for the production of official statistics, will be made available at the time of release to provide advice on data, to comment on interpretation and fitness for purpose, and to ensure the release process is correctly completed. Their name(s) and contact details will be clearly displayed on the statistical output. As set out in the Civil Service Code, any contact with the media must be agreed by ministers.

Statistics will be released as soon as practicable once they and any accompanying commentary or analyses are judged fit for purpose. There must be no opportunity, or perception for opportunity, for the release of unfavourable data to be withheld or delayed for political reasons.

An exception may be made where two sets of figures measure facets of the same issue, and simultaneous release would present a more coherent picture to users. In this case the release of one set of statistics may be delayed to coincide with publication of the other, provided users' wishes are taken into account. Such decisions on ‘bundling’ of releases will be taken by the Chief Statistician in the Welsh Government.

For periodical releases, the decision on timing will usually involve a trade-off between achieving the earliest possible release for a particular period, and maintaining consistency over time. Such decisions will be taken by the Chief Statistician in the Welsh Government.

For complex or high profile releases statisticians may arrange a technical briefing for the media at the time of release to aid interpretation of the findings.

In the event of technical issues relating to publication on the main Welsh Government Statistics & Research website, publications can be released through StatsWales and users notified as appropriate through social media and email.

Published information to accompany statistics

The release of any official statistics will be separate from statements that include presentation or advocacy of any related policies.

There will be occasions when statistics and a separate policy statement will be released at the same time. Although they may be released simultaneously, it is important that policy statements are not released in advance, as they should be seen to be a response to the statistics.

Where it will help users to understand the data contained within a release, a factual statement of the policy context will be included. This must do no more than state a policy objective and must never contain anything which could be interpreted as political comment. Such contextual policy statements must be used consistently over time and cannot be inserted only in those periods when they might be judged to have political value. Responsibility for wording lies ultimately with the Chief Statistician in the Welsh Government.

All regular statistical outputs will be released on the internet, free of charge.

In addition, we also publish data in response to user requests for additional analyses. We do not generally charge for provision of such data, however where demands were particularly large we would consider provision/costs in line with Welsh Government Freedom of Information guidance.

The Chief Statistician in the Welsh Government will report regularly on compliance with pre-release arrangements via the Welsh Government internet site.

Statistical Services in the Welsh Government will maintain and publish a statement describing how it applies these standards to each of its releases.

Any details of exemption to the Code of Practice will be published, as agreed with the UK Statistics Authority.

Any non-compliance will be reported immediately to the Chief Statistician, who will then decide the necessary action needed. Breaches of the Code of Practice for Statistics will be reported to the UK Statistics Authority and published on the Authority and Welsh Government websites.