jpr / policy paper No. 2 1999
Ethnic and religious questions in the 2001 UK Census of Population policy recommendations
'The national census can be an important means of fostering a multicultural society and a participatory democracy. Redesigning it for this purpose can have long-term social, political and economic benefits for British society.'
Barry Kosmin
Summary of key principles and recommendations
The value of the census in a multicultural society
- The national census can be an important means of fostering a multicultural society and a participatory democracy if it is redesigned for this purpose.
- The census questionnaire must be seen as an educational and social policy instrument which requires careful construction.
- The collection of data on minority groups can assist in combating and monitoring discrimination and prejudice, and also help minorities to develop their community organizations and services.
The following specific recommendations are made in the light of these principles:
Religion question
1 A religion question should be included in the census.
2 Response should be voluntary.
3 'Jewish' should be a recorded response.
4 A variety of Christian denominations and churches should be recorded.
Ethnic question
5 The ethnic question should be expanded to record a variety of 'white' ethnic/ancestry groups.
6 'English', 'Scots', 'Welsh', as well as 'Irish' , categories should be recorded throughout the UK.
7 'Jewish' should be a recorded ethnic category.
Question format
8 The question format should avoid any appearance of hierarchy or order of preferred answers/categories, especially by a vertical listing or numbering system.
9 The ethnic question should avoid offering 'mixed race or ancestry' as a 'bottom of the list' option.
10 The options and categories for both census questions should appear in random order and adopt an 'open' response or a write-in system.
11 The government should adopt the approach and format of the Census of Canada on ethnic and religious questions since they meet all the criteria outlined in this paper.
12 Adopting the 'user friendly' Canadian system will increase participation rates among minority communities. Aside from social policy benefits, this will reduce the under-enumeration problem. The Office of National Statistics (ONS) should transfer the resulting administrative cost savings to data capture and processing so as to allow for the extra work in the coding of responses which is associated with the approach advocated in this paper.
Introduction: devising a national census for a participatory citizenry in a multicultural society
Once every ten years since 1801 there has been an official count of the inhabitants of Britain through the national census. The decennial census is a compulsory activity completed by every household in the country and the collated results form the core of the national statistics. The content, production and release of the results are formally overseen by Parliament. The essence of the census is its universal coverage, and as a source of local area statistics it is of unique value.
Yet over recent decades the conception of the census has changed. Originally it was a method of understanding the country, its composition, its peoples and the activities taking place. Now its purpose is to provide information to central and local government, the health service and so on. But whilst the census appears to be simply a pragmatically-designed instrument for collecting policy-relevant statistical data, it draws on unexamined assumptions and may have unintended political consequences. The government should re-examine the purpose of the census, especially as it relates to efforts to foster a multicultural society. The aim of this paper is to set out the case for facing this issue head on by devising census questions and classifications that enhance social cohesion through respecting and honouring difference and diversity.
Unfortunately, rates of public participation and involvement are in decline in many areas of life, such as voting and donating blood. In recent decades many western democracies (such as the Netherlands, Germany and the USA) have faced problems in carrying out their national census due to a decreasing level of co-operation from the public. This has resulted in a growing inability to obtain full coverage of the population. Known as the 'undercount', this problem has largely been seen as one of technical 'data capture' and the domain of social statisticians and survey experts. It is, however, primarily a political and social policy issue and resolving it would be an important and positive development for civil society.
The census as 'civic ceremonial'
An emphasis on the civic ceremonial of the census would benefit the body politic. Even more than a general election, a national census allows the whole population to engage in a common activity for a common purpose. It should be an important symbolic act that strengthens feelings of inclusion and a general sense of community and trust throughout the country.
Regrettably, the trend in society is not in this direction. In most western countries the 'undercount' is a feature of the inner city where there is a particularly high concentration of the socially alienated members of society. Minorities and the socially and economically disadvantaged are generally sceptical of statistics and they have good reasons for their traditional hostility to government data-gathering. Historically the national census has long been an exercise in power and compulsion. The Latin word statisticus means 'of state affairs'. For centuries statistics were gathered by rulers and governments for purposes of taxation and military conscription, which bore heavily on minorities. For much of this century the census was used in the states of central and eastern Europe to create the basis for educational and job quotas (numerus clausus) detrimental to Jews and other minorities.
In the less fearful and deferential atmosphere of contemporary society distrust of government and scepticism about the motives of the census takers and the use they make of the results is no longer confined to the socially excluded and inner city dwellers. It is now quite widespread among many sections of the general public. The public mood is to resent the form-filling as an imposition and a burden which brings them or their families no apparent or tangible benefit. This is not an argument for abolition of the census but there is obviously a credibility gap. There is therefore a great need to 'sell' the census as a worthwhile common national activity. Given the right atmospherics, the whole population engaged in something for the collective good on a single day could be transformed from an imposition into a symbolic act of solidarity and common citizenship.
In contrast to much of the population, those in positions of power in society-the decision-makers in national and local government and especially big business-are enthusiastic supporters of such exercises in national data gathering. They have (or think they have) a great need to know how many bedrooms you have, how long it takes you to travel to work, how many GCSE subjects you passed and what sorts of people live with you. Companies use census data to make all manner of decisions. Developers rely on the number-of-bedrooms question to determine what size homes to build in a community. Department stores and supermarkets use data on working mothers to decide how to stock their shelves. This kind of detail is much sought after by industry and retailers, and with government doing the collecting, it is free. Meanwhile the public pays twice: once in taxes and then again in time spent around the dining-room table providing this marketing-rich data.
It is important for national social policy that the sponsors of the census recognize that there is a real need to consider its potential impact on society. When the Office of National Statistics (ONS) speaks of 'a demand-led census', there ought to be an acknowledged obligation to accommodate the needs of communities and community organizations as well as business and public authorities. This obligation is especially important in the area of the identification and categorization of minorities-the ethnic and religious questions in the census. Handled wisely these questions could enhance community and race relations and a sense of belonging. This in turn could help reduce the undercount in inner city areas and so the overall cost.
To understand the significance of the census for community relations, its informational and educational impact must be appreciated. The census is a much heralded event that requires universal participation. Millions of people who would otherwise not consider such issues have to decide how to answer demographic, educational, employment and housing questions about themselves and all the members of their household. The content and wording of the questions inform and educate the British public as to what the government thinks it ought to know about everyone, as well as how it conceptualizes issues and groups in society. Thus the census establishes 'the official cognitive system' and it transmits that message into every home in every street and hamlet in the nation.
In no area of enquiry is the role of the national census more important in educating the public and establishing a discourse for imparting this information than in questions about cultural identity. The ethnic question creates categories and identifies communities which affect general usage in society as a result of their take-up by the communications and media industry. What is institutionalized in this process is a vision of society and its make-up: who the minorities are and who the majority is and on what basis these groupings are defined. Whether one likes it or not, the census is not just about collecting statistics and tabulating them but also about setting social policy agendas.
The following policy recommendations on the collection of religious and ethnic group data for implementation in the 2001 Census are made in the light of these arguments and concerns.
Policy recommendations: the rationale
Religion question
1 A religion question should be included in the census
The case for a religion question is essentially practical. Religion is an important organizing principle in British society and an important element in the lives of many citizens. Religious groups and communities need census-type data on their members and potential constituency in order to operate and plan more efficiently for the myriad of educational, welfare, social and recreational activities and organizations that are presently run under religious auspices.
2 Response should be voluntary
Again the reason for this is practicality. The introduction of a religion question will require changes in the 1922 Census Act. Undoubtedly there will be opposition to the change on grounds of conscience and personal privacy. This will probably extend to completion of the census questionnaire. Most Commonwealth states that have a religion question in the census (Australia, Canada, New Zealand, South Africa) disarm this opposition by making response to the question voluntary. Their experience suggests that the normal rate of non-response on grounds of conscience is under 20 per cent. Thus useful data will still be collected on the vast majority of the population while boycotts and civil disobedience can be avoided.
In this regard the use of the term 'faith' in any question should also be avoided. The form that should be adopted is the simple question 'What is your religion?', with a note to make it quite clear that the enquiry seeks to enumerate those with a specific religious preference.
3 'Jewish' should be a recorded response
Census data on the Jewish population, particularly at the level of small area units, will be very valuable to planning for the future of the Jewish voluntary sector. As the Board of Deputies of British Jews has stated, the practical benefits arising from the collection of this data outweigh the dangers of its theoretical misuse by any future government. There is also widespread support for this initiative across the spectrum of religious communities-a significant measure of the health of British civil society and of trust in government.
4 A variety of Christian denominations and churches should be recorded
The various branches of Christianity operate separately in areas such as education and social provision. Therefore, aside from difficulties of mutual recognition, there are solid practical reasons for their numbers and distribution to be recorded separately (as in Northern Ireland). The idea that there should be a 'catch-all' Christian category covering the vast majority of the population followed by a large variety of small and 'exotic' alternatives should be strongly opposed. This approach would create an inaccurate picture of a totally dominant majority religious group and a large number of minor and insignificant others. It would also be very detrimental to community relations and the support of minorities for this data-gathering exercise.
Ethnic question
5 The ethnic question should be expanded to record a variety of 'white' ethnic/ancestry groups
The introduction of an ethnic question in the 1991 census was a positive move for British race relations in terms of protecting minorities from prejudice and discrimination. However, it had the detrimental effect of creating a false notion of homogeneity, referred to above, whereby 95 per cent of the British public were placed in an artificial racial category called 'white' and the 'non-white' remainder was placed in a variety of categories. Theoretically and politically this crude 'colour coding' approach to the ethnic question is wrong, but there is nevertheless a need for ONS to replicate a new census question in order to provide for a time series.
Therefore, the 'white' category should be expanded to allow for more specific responses on ancestry as indicated below. This will have the benefit of reflecting social reality and help remove the false impression given to the public that only 'non-whites' have ethnic identities and ties.
6 'English', 'Scots', 'Welsh', as well as 'Irish', categories should be recorded throughout the UK
This recommendation obviously assists the aim of recommendation 5 above to de-emphasize the 'white' racial category. Irish identity should be separately recorded, as requested by the Irish community and the Commission for Racial Equality (CRE).
In addition, there is a related concern in this area regarding devolution, the political and constitutional significance of which is still little appreciated. Given their historical experiences, Jews are naturally sensitive to the dangers of an increased emphasis on national identity which the introduction of legislatures in Scotland and Wales might bring, and of any encouragement of exclusive nationalisms based upon false notions of ethnic homogeneity in these countries.
Constitutionally Scotland and Wales are civic nations. Residents in these countries will be able to vote for the assemblies irrespective of their ethnic backgrounds or place of birth. Since no local citizenship is involved ethnic Scots living outside Scotland have no voting rights for the Scottish Parliament. These facts need to be publicly recorded and communicated and UK-wide census results can be an effective medium for this purpose. The collection of ethnic data will be a symbolic appreciation of the existence of minorities and pluralism in the devolved nations as well as in England.
7 'Jewish' should be a recorded ethnic category
Since Jews are recognized as an ethnic minority under the Race Relations Act this would be a logical move. Modern European antisemitism has been based on racial and biological theories rather than on anti-Judaism. The targets and victims of antisemites are people belonging to the Jewish 'racial/ethnic entity', as the antisemites define it, including people of Jewish ancestry who follow faiths other than Judaism or none.
In addition, the Jewish community is more than simply a religion in the conventional Christian sense of 'a community of faith'. It has ethnic and communal dimensions that are both part of and stand somewhat separate from the Jewish religion. Thus many Jews who reject a religious self-definition will identify on the ethnic criterion. This societal reality is confirmed by the Canadian census, where 'Jewish' is recorded in both the religious and ethnic ancestry questions.
Canadian Jewry is comparable in size, origin and structure to British Jewry so the results of the Census of Canada are significant evidence of the need to record Jewish identity in two ways. In Canada in 1991 the total population that self-identified as Jewish on either question was 405,955 persons. The religion question yielded a population of only 318,070, while 281,680 self-identified as Jewish on both the religious and ethnic question.
There are thus external and internal reasons for collecting both ethnic and religious data on Jews. Only in this dual way can a full count of all those who wish to self-identify as Jews emerge.
The structure and format of census questions
8 Avoid hierarchical ordering
An appreciation of the need for delicacy in handling ethnic and religious questions in the census has already been stressed. Cost-saving devices that would ease the burden on ONS should not be adopted if they hazard community relations. It would not be wise to list categories or options for either census question in any way that might be interpreted as a hierarchy of preferred answers, for example by vertical ordering or a numbered list.
9 Avoid 'mixed race or ancestry' as a 'bottom of the list' option
This concern is particularly relevant in the case of groups who are regarded as 'problem cases' by the technical experts on data capture grounds. In particular the ethnic question must avoid offering a 'mixed race or ancestry' response as a 'bottom of the list' option, which could perpetuate stigmatization of what were once unfortunately known as 'half castes'.
10 Adopt an 'open' response format
The best solution to this challenge is to adopt an 'open' response format whereby the respondent inserts his/her own answer after reading an alphabetical list of the main possibilities. This will also dispense with any objection that the people are being restricted in their responses by an official classification system which wishes to impose a state-sanctioned system of self-identification. ONS objections to this approach on grounds of practicability and data processing can be largely countered by the success of the US Bureau of Census with the 1990 US Census, which offered respondents a choice of up to four separate 'write-in' answers for the ancestry question.
11 Adopt the model of the Census of Canada
Canada has had ethnic and religious questions in its national censuses throughout the twentieth century. This fellow Commonwealth parliamentary democracy also passed a 1988 Act for the Preservation and Enhancement of Multiculturalism. Canada provides a good role model for the UK to follow on technical as well as political grounds.
The government should therefore direct ONS to follow the procedures of the 1991 Census of Canada question on religion (see appendix) and the 1996 ethnic origin question.
12 Cost savings should free resources for extra coding
Following the Canadian system will have the positive effect of increasing the response rate among minorities and so reduce the 'undercount' and the administrative costs associated with under-remuneration in the census process. These cost savings can then be transferred by ONS to the data capture and data processing procedures to allow for the extra coding which the approach advocated in this paper involves.
Conclusion
Encouraging minority communities to participate in the census as citizens and making it possible for them voluntarily to identify themselves in their own terms can be an important means of fostering a multicultural society. This approach can have long-term social, political and economic benefits for British society.
Appendix
Report author
Professor Barry Kosminis JPR's Director of Research. He has written widely on issues concerning contemporary ethnicity and religion having served as a Consultant to the SSRC/ESRC Research Unit on Ethnic Relations at the Universities of Bristol and Aston.
Based in the USA from 1986 to 1996, he directed the City University of New York's National Survey of Religious Identification, the largest ever survey of American religion covering 114,000 households. Professor Kosmin was also Director of the 1990 National Jewish Population Survey for the Council of Jewish Federations.