Last updated: July 2011
This policy will be reviewed initially after six months of implementation and thereafter on an annual basis.
Next review will take place in July 2013
1.1 Statistics are by their nature subject to error and uncertainty. Initial estimates are often systematically amended to reflect more complete information. The Office for National Statistics (ONS) will ensure that revisions to its published statistics are accurate, robust and freely available and that they are accompanied by the same supporting and explanatory information as are the statistics.
2.1 ONS will carry out its responsibilities in respect of revisions by following this policy, which will ensure it is performed to the requirements of the National Statistics Code of Practice.
2.2 A revision (R) is a scheduled change to any published ONS output whether printed or electronic. Revisions are categorised as follows:
R1. To incorporate better source data such as:
more complete or corrected reporting by respondents
source data that more closely match the concepts
R2. To capture routine recalculations, such as:
Incorporation of updated seasonal factors
Updating of the base period
R3. To reflect improved methodology:
Changes in statistical methods
Changes in concepts, definitions and classifications
2.3 A correction (C) is an unplanned change to any published ONS output whether printed or electronic.
2.4 A labelling convention is used as follows:
R1 and R2 are 'Updates'
R3 are "Improvements'
C are 'Corrections'
3.1 ONS will publish and maintain a general statement describing its practice on revisions. This statement will describe the overall approach to revisions for R1, R2 and R3, emphasising the need for revisions to be scheduled to avoid unnecessarily frequent revisions
3.2 Key outputs, subject to scheduled revisions, will have a published revisions statement. The revision statement will set out the general schedule for revisions, and take into account the ways in which the output is used, the resources available for its production. It should be subject to user consultation.
3.3 Revisions will be released in compliance with the same principles for other new information. This will cover standards in the release practice and should also consider arrangements for early access.
3.4 ONS will announce substantial methodological changes (R3) before the release of statistics based on the new methods. In the case of key outputs a information a paper should be made available via the website.
3.5 ONS will minimise the possibility of corrections (C). If they do occur they will be released promptly and in an open and transparent manner, and, where material, stakeholders alerted.
3.6 Major revisions and corrections (for any reason, that is, R1-R3, C) will be accompanied at the time of publication by an explanation of their nature and extent. Updated back data will be provided if available and where it will add value.
3.7 ONS will monitor the long term effects of revisions on key outputs. This will be done with a view to improving the quality of these outputs.
3.8 In planning revisions due to improvements (R3), ONS will consider carefully the balance between 'bunching' revisions versus many successive revisions, in consultation with users. Whilst users often prefer bunching to avoid multiple revisions, this does lead to larger, more complex projects and difficulties in distinguishing separate effects.