On 18 January 2012 the Economic and Social Data Service released for secondary analysis the dataset from the ‘ONS Opinions Survey, Census Religion Module, April, May, June and July, 2009’. This is available, under special licence access to approved UK researchers (accredited by the UK Statistics Authority), as SN 6938. For further information, see:
The Office for National Statistics (ONS) Opinions Survey (OPN), previously known as the ONS Omnibus Survey, is a regular, multi-purpose study carried out by the ONS Social Survey Division. It started operating commercially in 1990 and was conducted for eight months of the year until April 2005 and monthly thereafter.
A census religion module (MCG/MCGb) was included in the OPN for April-July 2009 inclusive, as part of the final testing of question-wording for the 2011 population census. Citizenship was also covered in the same module (in April and May). A total of 4,235 Britons aged 16 and over living in private households were interviewed face-to-face.
The question which was tested on religion is one which is not often used in sample surveys. ‘Which of these best describes you?’ was followed by eight reply options: Christian, Buddhist, Hindu, Jewish, Muslim, Sikh, any other religion (specify), and no religion. In the May and July 2009 surveys no religion was made the first option, ahead of Christian. Any spontaneous comments made by the respondent to the question were also captured by the interviewer.
As well as through the OPN, ‘Which of these best describes you?’ was evaluated through: a postal test in England in March 2009 (with no religion given as the first option), cognitive testing, and engagement with key stakeholders. For comparative purposes, another question – ‘What is your religion, even if not currently practising?’ – was included in the core questionnaire for the April-July 2009 OPN.
In the end, ONS decided against using ‘Which of these best describes you?’ in the 2011 census and in favour of ‘What is your religion?’ – which many commentators regard as potentially leading. The ONS rationale for doing so is set out in the October 2009 report Final Recommended Questions for the 2011 Census in England and Wales: Religion, which is available through the Government web archive at:
In Annex A of this document ONS tabulated the results from the core and module questions on religion in the April-July 2009 OPN. ‘Which of these best describes you?’ was found to increase the proportion professing no religion compared with ‘What is your religion, even if not currently practising?’ But the difference was especially noticeable in May and July, when no religion headed the list of options. In this instance, perhaps it was the running order of options more than the question-wording per se which most affected the results.
So, these April-July 2009 OPN data do not simply have historical significance. They remain important methodologically in demonstrating how variations in questionnaire design can impact upon the statistics generated by enquiries into religious affiliation. Doubtless, the first results from the religion question in the 2011 census, when they come, will reignite the debate about what is the ‘right’ way to formulate this question.