Freedom response to religion question 2021 census

​FOI REF: FOI/2022/4663

You asked

In your publication today on religion in the 2021 UK census, you specify a total of 58 categories, as shown at:

https://www.ons.gov.uk/datasets/TS031/editions/2021/versions/1

Not included in these categories is the religion "freedom". Accordingly, I am writing to request the following information:

1. the number of responses to the 2021 census in England & Wales identifying their religion as "freedom", "Freedom", or "FREEDOM";

2. your methodology for coding such responses (for example, are they coded under "Other religion: Other religions"?); and

3. all correspondence (whether internal or external), memos, reports, research, and other documentation informing your decision on how to code such responses.

We said

Thank you for your Freedom of Information request regarding write-in answers of 'Freedom' (or variations thereof) to the religion question in Census 2021.

On 29 November, we published our topic summary and data on religion from Census 2021, which includes statistics and commentary on people's religious affiliation across England and Wales. Religious affiliation is a measure of how a person connects or identifies with a religion, whether or not they practise or have belief in it.

As you note in your email, the Religion (detailed) dataset includes estimates for 57 answers to the religion question. 'Freedom' was not included as an individual category among the 5 non-religious categories in this dataset.

From phase 3 of our Census 2021 release plans, we can produce write-in answer counts for 'Freedom' via our bespoke servcies. Such services would be subject to legal frameworks, disclosure controls, resources and agreement of costs. Bespoke requests can be made via our Census Customer Services team Census.Customer.Services@ons.gov.uk.

In answer to your second question, the hundreds of thousands of write-in answers to the religion question were sense and spelling checked and then grouped. The groups included in our published datasets are not based on numbers alone; review and development included:

• new user needs (identified through stakeholder engagement and a public consultation)
• the groups published for the 2011 census data (to help with consistency for comparing data between censuses)
• the number of people who responded with the same identity (to ensure confidentiality in our published data

This included a public consultation last summer on our content and release phase proposals for Census 2021. In this consultation we asked for feedback on whether our proposed table specifications for our topic summaries met user needs. We did not receive any requests for the inclusion of 'Freedom' (or variations of it) in our religion (detailed) topic summary data tables. Further information on this can be found in our Response to the 'Census 2021 outputs: content design and release phase proposals' consultation – Part one report.

For further information on how write-in responses are included in our classifications, please see our blog post, How am I represented in Census 2021 data?

Everyone is represented in our statistics, but not every write-in answer is published. If a certain answer is not included in the published dataset individually, it will be grouped with other related answers in the published data. People who wrote in 'Freedom' were included in the 'No religion: No religion' estimates in the Religion (detailed) dataset.

In response to your third request, we have also included two emails and a spreadsheet (which was attached to the second email). This documentation relates to the Census 2021 'live operations' write-in answer coding process. During this period our automatic coding system provided 'residual' write-in answers that could not be categorised automatically (via spell checks or word lists most often).

The emails refer to the meeting in which 'Freedom' was discussed and the spreadsheet indicates the decision to add it to the automatic coding word list for the 'No religion' category. This decision was made to ensure all 'Freedom' write-in answers (or variations thereof) were included in our statistics. It is also the reason that we may be able to produce a write-in count, as previously mentioned.

Personal data has been redacted under Section 40(2) of the Freedom of Information Act 2000 (FOIA).