Data and methods that produce assured statistics

Quality means that our outputs are fit for their intended uses, are based on appropriate data and methods, and are not misleading.


The census only happens every 10 years and counts every person and household in England and Wales. The census is unique. There is simply nothing else that gives so much detail about us and the society we live in. The census asks everyone in England and Wales the same questions about the same day. This makes it easy to compare different areas and populations.

We try our hardest to make sure we count everyone, however, no census is perfect and it can miss a number of people or count some more than once. The Office for National Statistics (ONS) will use complex statistical techniques to adjust Census 2021 counts for those people missed by the census.

In October 2020, we published our statistical design for Census 2021 in England and Wales. This end-to-end statistical design has been created to ensure Census 2021 results are of high quality and fit for purpose. The statistical design focuses on four main phases:

  • design and build

  • monitor and counter

  • process and estimate

  • outputs and disclosure control 

In 2015, we established an external Methodology Assurance Review Panel (MARP) to:

  • provide external, independent assurance and guidance on the statistical methodology underpinning Census 2021 estimates and those based on administrative sources

  • identify significant gaps and risks in methods and make suggestions for mitigation

  • review admin data methods and contribute to their continuous improvement

A list of methodology assurance papers submitted to MARP have been published and represent the ONS's methodology at the time of each submittal. A selection of these include:

More information

If you would like to find out more about Census 2021 quality and methods, please email us at with your queries.