1. Introduction

Changes or updates to published statistics, analysis or data are sometimes required for various reasons. We use different terminology depending on the reason for the change or update. These terms are “revisions” and “correction of errors” and are defined in Section 3: Revisions Policy, and Section 4: Correction of Errors Policy.

Being clear, transparent, and consistent when publishing revisions and corrections of errors is important as it allows users to understand how and why our statistics, analysis or data are revised or corrected after publication, and informs their appropriate use. This approach also ensures that we are proactively meeting the Code of Practice for Statistics.

2. How to use these policies

These policies set out definitions of revisions and correction of errors to provide coherent central changes and updates policies for all Office for National Statistics (ONS) statistics, analysis, and data. These unified policies enable users to obtain information on terms used when reporting on revisions and correction of errors. This includes definitions that are used when referring to statistic or data changes and updates.

The Revisions Policy and Correction of Errors Policy have been separated on this page to clearly highlight the differences between them and to demonstrate when each should be used. This separation is standard practice used by other National Statistical Institutes (NSIs).

3. Revisions Policy

Revisions are updates to previously published statistics, analysis or data that improve quality by incorporating improved methods, additional data sources or statistics that were unavailable at the point of initial publication. Revisions should not be confused with the correction of errors; revisions are not published to correct a mistake.

We aim to limit the impact and number of revisions by finding an appropriate balance of data quality trade-offs, in particular between timeliness and accuracy. We produce initial estimates to meet a strong user need. However, this means that there are situations in which revisions are required to ensure data quality and to meet the Code of Practice for Statistics’ pillars of Trustworthiness, Quality and Value. These situations may arise when further data become available because of:

  • incorporating data to estimates that were not available at the time of the initial publication, which includes replacing projections with actual data

  • benchmarking – some short-term, in-year statistics may be produced using data sources that are readily available but of a lower quality; when higher quality data sources become available, short-term statistics can be benchmarked against them and appropriate adjustments made

Situations in which revisions are required may also arise because of changes to methods or systems, including:

  • new methods or improved methodology, such as changes in statistical methods, techniques, or systems; this also includes changes in concepts, definitions, and classifications
  • rebasing – when a statistical index is rebased, it can be because either individual component items have been re-evaluated and the “weight” attributed to each has changed, or the reference period has been updated

Revisions are usually reported in the next release of the data by including a notice within the latest release. The timing that these are released will consider the need to:

  • make the information available to users as soon as is practicable
  • avoid two or more amendments (to the same data items) in quick succession

4. Correction of Errors Policy

Despite having quality control measures in place, mistakes may be found in our statistics. Correction of errors are amendments that are made to published statistics in response to the identification of mistakes following their initial publication. Corrections should not be confused with revisions; corrections of errors are not published to incorporate improved methods, additional data sources or statistics that were unavailable at the point of initial publication.

Correcting errors should be viewed as an improvement to data quality, to ensure that the best possible statistics, analysis, and data are available to our users. They are also an important part of transparency, which helps to sustain public confidence.

Errors may be from not carrying out the calculations correctly, or where we have identified that a method we have applied was incorrect. We may also find that we need to correct our statistics because there were undetected errors in the input data used to calculate the statistics.

When a mistake is identified, the error is corrected as soon as possible in line with the Code of Practice for Statistics, with limited exceptions. In a small number of cases, minor errors might not be corrected because of:

  • the time scale (for example, the error occurred too long ago)

  • the scale of the error (meaning that it would take too much resource for minimal improvements)

  • it not being in the public interest to make the correction

Correction of errors are reported at the top of the affected release in a blue collapsible banner titled “Correction”. This acts as a note on the publication that an error has been identified and includes details on the nature and rationale for it not being corrected. The fact that the publication has been updated is clearly stated with details of when and why the change was made.

5. Links to current revisions and correction of errors policies

In some cases, Office for National Statistics (ONS) publications have specific revisions or correction policies that describe details for users on how revisions or corrections are managed. The list below provides links to currently available publications specific policies. These complement the policies detailed in Section 3 and Section 4.

Economic statistics

Our Revisions policies for economic statistics web page includes links to the:

  • Investment by Insurance Companies, Pension Funds and Trusts (MQ5) Revisions Policy
  • Public Sector Statistics Revisions Policy and Practice
  • Public Service Productivity Estimates Revisions Policy
  • Retail Sales Index Revisions Policy
  • Wealth and Assets Survey (WAS) Revisions Policy

Labour market

Our Revisions policies for labour market statistics web page includes links to the:

  • Labour Market Statistics Revisions Policy
  • Public Sector Employment Statistics Revisions Policy

Population and international migration

Business surveys