Introduction

The Office for Statistics Regulation (OSR) is the regulatory arm of the UK Statistics Authority. Under the Statistics and Regulation Service Act 2007, OSR independently reviews whether official statistics comply with the standards of Trustworthiness, Quality and Value in the Code of Practice for Statistics (the Code).

We do this through a tailored process called a compliance review, where we evaluate the extent to which official statistics meet the standards set out in the Code.

Types of compliance review

A compliance review is the umbrella term for a review of any official statistics against the Code carried out by OSR.

There are three types of compliance reviews:

• Assessment: a review conducted when a producer has requested the accreditation of official statistics. Compliance with all standards and practices of the Code’s Standards for Official Statistics is reviewed. Assessment is the only tool that allows us to grant accredited official statistics status (called National Statistics in the Act).
• Compliance review of existing accredited official statistics: a review conducted to check the statistics’ ongoing compliance with the Code. They often focus on a specific issue or part of the Code. They may lead to more-focused reviews if issues arise and could ultimately result in us deciding to remove the accreditation from the statistics.
• Compliance review of official statistics: a review conducted to check the statistics’ ongoing compliance with Code, but where the outcome does not involve a decision on accreditation. Their main aim is to recommend improvements to the statistics. We may recommend that the producer seeks accreditation, but it is up to them to pursue accreditation separately via assessment.

Outcomes of compliance reviews

The potential outcomes of each type of review are different.

Where official statistics have been put forward for assessment, there are two potential outcomes:

• Award accreditation – we find that the statistics meet the standards of the Code. We award the accredited official statistics status after assessing the statistics, either straight away or once some requirements have been met.
• Do not award accreditation – we find that the statistics do not meet the standards of the Code. We do not award the accredited official statistics status.

For a compliance review of existing accredited official statistics, there are two potential outcomes:

• Reconfirm accreditation – we find that the statistics still meet the standards of the Code and can continue to be published as accredited official statistics. We may identify recommendations for improvement.
• Remove accreditation – we find that the statistics no longer meet the standards of the Code and should be published only as official statistics.

For a compliance review of official statistics that is not an assessment, there is only one outcome: a summary of areas of positive work and recommendations for improvement.

Outputs from compliance reviews

All assessments, and most other large or complex compliance reviews, will have two outputs:

• a report with key findings and requirements or recommendations; and
• an accompanying letter to the Head of Profession for Statistics (HoP), Chief Statistician or other senior statistician with a high-level summary of our findings and next steps.

Smaller compliance reviews will also have two outputs, but these will be shorter:

• an HTML page setting out the key strengths of the statistics, the main areas for improvement (recommendations) and next steps; and
• a short accompanying letter to the HoP, Chief Statistician or other senior statistician thanking the team for their time and directing them to the HTML page with the key findings.

Scope of compliance reviews

For assessments, the scope is always the same: we review compliance with each standard and practice of the Code’s Standards for Official Statistics.

For other compliance reviews, the scope is more flexible. The review may be a light-touch check of compliance with the TQV framework, or the focus could be a specific standard or core principle (Trustworthiness, Quality or Value). We decide which approach we take on a case-by-case basis. For instance, if there are known concerns about the quality of data sources and methods, then the review would likely have a quality focus. Typically, most reviews will consider some aspects of value such as the clarity and insight of the statistics. If we’re reviewing a set of statistics produced by a well-established producer whose statistical policies and practices we know well and which we have reviewed recently, there may be less focus on the trustworthiness standards. This approach allows us to focus our time and effort on the areas of greatest relevance and concern.