About the census : From collecting to publishing census information

To produce statistics, we need to turn answers from census questionnaires into data.

First, we compile the data. This means we collect and scan paper questionnaires and securely combine the information from them with the answers from the online questionnaires.

Next, we clean the data. That means we check for obvious errors, such as someone accidentally filling in both a paper and online questionnaire for the same address.

Using a recognised statistical method, we also estimate missing responses to mandatory questions.

These standard processes are used on most official statistics (opens in a new tab)  produced from surveys.

We carry out the Census Coverage Survey (CCS) to estimate how many people the census missed.

By combining the results of both the census and the CCS, we can use recognised statistical methods to get a full estimate of all the people in England and Wales.

These processes allow us to create a high-quality set of statistical data from which we produce statistics. These statistics help organisations plan and fund services we all need. We keep every questionnaire safe and confidential for 100 years, as we describe in Your data and security.

Our quality assurance begins as soon as the first census information arrives and continues throughout the processing steps. We look for unusual or unlikely patterns in the data and we check that our processes have worked as expected.

This work covers a wide range of checks and includes carrying out the Census Quality Survey. As detailed in our quality assurance plans on the ONS website (opens in a new tab) , we also use other data sources to cross-check census information. We protect that information in the same way as we protect census information. Find out about how we keep information secure.

Coronavirus (COVID-19) impacts

We recognise that the coronavirus pandemic had an impact on census information from some groups more than others. So, we’re giving special attention to our quality assurance of these groups.

For example, many students were unable to return to their term-time address after Christmas 2020. We took steps to help students complete the census for that address, using the benefit of having an online questionnaire. We’re doing extra work to quality-assure this information.

Using local knowledge

We invited every local authority to tell us about features of their area’s population that we could consider in our quality assurance. For the first time, we gave local authorities limited access to provisional estimates of the numbers of people in their authority. These estimates contained no personal information. The local authorities’ valuable local knowledge has helped us understand anything unexpected in the data. This access also helped us as part of our quality assurance process and, using strict security controls, was in line with the Code of Practice for Statistics (opens in a new tab) .

Find out more about our census quality assurance on the ONS website (opens in a new tab) .

We apply strong disclosure control protection to the data we plan to publish. Disclosure control is ensuring that statistics are useful and valuable for users, while protecting the confidentiality of information about individuals, households and organisations.

To further ensure the data meet the needs of census data users, we carried out a census outputs consultation.