Guidance

Statistical notice: MHCLG’s policy covering revisions to official statistics

Published 11 February 2010

Introduction

This statement describes the principles and practices to which analysts in the Department will adhere when handling changes to official statistics that have already been placed in the public domain - whether through a hardcopy document or electronically.

Our statement has been drafted in accordance with the Code of Practice for Statistics published by the UK Statistics Authority which acts as the regulator of the UK statistical system. It also conforms to the guidance issued to producers of official statistics by the Office for National Statistics entitled Communicating quality, uncertainty and change.

The UK Statistics Authority’s Code requires that scheduled revisions or unscheduled corrections to statistics should be released as soon as practicable, and that such changes should be handled transparently in line with a published policy.

For the purposes of the Code of Practice the Authority distinguishes these two categories of change as follows:

Scheduled revisions – Planned amendments to published statistics which are made in order to improve quality by incorporating additional or corrected data that were unavailable at the point of initial publication.

Unscheduled corrections – Amendments made to published statistics in response to the identification of inadvertent errors subsequent to their initial publication.

In the case of this Department we can add to the above two categories the occasional situation where, at the point of publication, statistics producers are aware of the need for a revision to (primarily quarterly) data but where the data are not finalised at that particular point.

Background

By their very nature, statistics are subject to uncertainty and error. It is often the case, for example, that statisticians need to amend their initial or provisional estimates in order to reflect the availability of more complete information, or where improvements in methodologies and underlying computation systems can help to make statistics more accurate and more useful.

Pre-planned changes such as these can be managed systematically, pre-announced, and reflected in publication schedules.

However, some changes to statistics will fall outside any systematic and managed pattern of revisions normally associated with a set of statistics because they are unexpected.

Such unanticipated changes to official statistics can be caused, for example, by administrative mishaps; by errors; by the unexpected acquisition of new data; or because of the effects of non-statistical changes to any underlying administrative or management systems.

The Department’s analysts endeavour to minimise the possibility of unexpected revisions but, if they occur, they will be handled in accordance with the management procedures described below.

In each of the circumstances described above, analysts will comply with the following broad principles:

Overarching principles

• We will maintain an overarching statement, such as this, which describes the way we handle both scheduled revisions and unscheduled corrections to our published official statistics.
• Changes to official statistics will be released as soon as practicable, and in an open and transparent manner. Our objective will be to maintain an appropriate balance between timeliness and accuracy.
• The publication of substantial changes^{[footnote 1]} to official statistics will be accompanied by a detailed explanation of their cause, nature, scale and impact.
• We will keep a historical record of any substantial changes to our official statistics, and the reasons for such changes, and use that record as a diagnostic tool in order to monitor and enhance the quality of our statistics.
• Ministers and officials will be given pre-release access to any substantial changes to our official statistics in accordance with the rules set out in the UK Government’s Pre-release Access to Official Statistics Order 2008.

MHCLG’s analysts will also adopt the following management practices:

Practices

In those circumstances where the Department’s official statistics are subject to a pre-planned and managed pattern of changes:^{[footnote 2]}

• We will publish a list showing the titles of any of our official statistics that are subject to scheduled revisions.
• All official statistics outputs that are subject to scheduled revisions will have their own dedicated and published policy which explains how such revisions will be handled.
• For each output, users will be provided with the timetable for any scheduled revisions so that they are prepared for them when they are due.
• The release of any scheduled revisions will be accompanied by a statement which explains their impact on the statistics affected.

In those circumstances where our already published statistics require unscheduled corrections:

• If the changes to published statistics are minor in the sense that they have little or no material effect on the statistics, or on the way that those statistics have been interpreted, we may defer publication of the correction until the next release is due for publication in order to avoid causing unnecessary inconvenience to users.
• Otherwise, unexpected changes to published official statistics will be released as soon as practicable.
• The release of unscheduled corrections will be accompanied by a statement which explains why the correction was necessary, and what impact it has had on the affected statistics.

1. Substantial changes, whether scheduled or unscheduled, are defined as those which have a material effect on the quality and thus reliability of published statistics because they change the story told by the statistics, and any interpretation that can be drawn from those statistics. ↩

2. Due to processing limitations, revisions are not applied to the Department’s monthly and quarterly statistics on Local Authority Borrowing. Revised data are, however, submitted to the Office for National Statistics. ↩