Data collection and publication - religion/belief: guidance

Additional guidance for interviewers

Instructions and 'prefer not to say'

Instructions should be given at the beginning of the interview/survey saying why all of the questions are being asked, that they are all voluntary and, if a respondent does not wish to answer any of the questions, they do not have to. This ensures all questions are treated the same.

Testing by the ONS has shown that provision of a 'prefer not to say' option results in an increase in non-responses in telephone/face-to-face interviews. As such the recommendation is that this should only be included in online and paper surveys where refusal is not otherwise possible with the caveat that this may increase the rate of non-response to this question.

If a 'prefer not to say' option is added to this question then it should be added to all questions. In interviewer-led surveys clear instruction at the beginning should advise people that they do not have to answer any question that they do not wish to.

Presentation of the questions

It is strongly recommended that you do not change the wording of the question as it could have implications on the data you receive. The wording of this question affects the way in which people answer, especially if they only have a loose affiliation with religion. Any changes to the wording of the question may result in people either not answering or providing a less suitable response and you will not be able to make comparisons to results from the 2022 Census.

The categories for the recommended question are the same as the 2022 Census, which have undergone minor changes since the 2011 Census. These are:

• the addition of the 'Pagan' category (which was included the recommended question in the previous version of this guidance)
• the addition of a write in option under the 'Muslim' checkbox
• the change in position of 'Hindu', from 9^th to 6^th

The categories have been carefully selected to best represent the Scottish population and to ensure the highest response rates by making it relevant and acceptable to respondents. This list of categories is not definitive and a write in box must be provided so that respondents can answer as they wish. Making changes to the categories could affect response rates if people feel that they are unable to record their religion accurately. It is very important to the completeness and acceptability of the question that the 'Other' write in box is retained. If you need to make changes to the categories, for example you require information on a specific religious group not listed, please contact us directly for advice. It is recommended the breakdown of the Christian category is used when asking this question in Scotland in order to facilitate more detailed analysis of this group.

Multiple ticks

The recommended religion question is a single tick response question. In order to obtain a complete count if a respondent has provided multiple responses then the rule should be that the response is recorded as 'don't know'. The exception to this rule would be if a respondent has ticked one of the categories and also the 'Other' box. In which case:

• if the respondent has ticked a category and written in a response in the 'Other' write in box then the write in answer should be used
• if the respondent has ticked one category and then ticked the other box without writing in a response then the first category they ticked should be used

Proxy responses

This question is designed to collect information on a person's own religion and should be addressed directly to the respondent. Where the respondent is unable to provide an answer then another member of the household, or the person's carer, can be asked to give a response on their behalf. This should be a last resort and the interviewers should record that a proxy response was given. It is very important that interviewers, administrative staff or anyone other than the individual's family member or carer does not provide an answer on behalf of the respondent.