1333. Information on the ethnocultural characteristics of the population is relevant for developing migration, integration, equality and minority policies. Countries may wish to collect information on the distribution of ethnic identities among the population, first or main languages, the knowledge and practice of languages, and religious affiliation. The combined collection and analysis of data on several ethnocultural characteristics may be particularly informative for the understanding of a country’s cultural diversity.

1334. All of the topics covered in this chapter are non-core topics. There are several reasons for this, which also constitute important considerations for any country collecting such information in a census:

(a) Some ethnocultural characteristics are politically sensitive. In some countries, collection of information on these topics in official statistics is prohibited by law.

(b) Ethnocultural characteristics may be subjective, or have a subjective dimension. Countries will adopt different concepts and definitions according to the prevailing understanding in each country, as well as to fulfil national policy needs.

(c) Countries using registers or other administrative sources for all or part of their census may be unable to use the census as a means of providing information on these characteristics, or may face constraints in the concepts and definitions used. For example, they may be limited to compiling information on formal membership of a church or religious community, or on the official language of communication used between the government and a household.

(d) Some ethnocultural characteristics identify very small populations, and members of minority groups may be vulnerable to discrimination. Special care may therefore be required in designing census procedures and outputs relating to ethnocultural characteristics, in order to reassure the public that appropriate data protection and disclosure control measures are in place.

1335. The right of respondents to declare their identity freely and openly is essential. The Human Rights-Based Approach to Data (HRBAD)¹³² stresses that this right to self-identification is a crucial element in data gathering and that it should be given high priority to the greatest extent possible, when dealing with subjective, dynamic or sensitive topics (such as ethnicity, indigenous status, religion, language, gender. As essential aspect of this principle is that individuals must have the option to disclose, or to withhold, sensitive information about their personal characteristics, and data collection activities should be conducted in accordance with the human rights principle of ‘doing no harm’. As such, countries may wish, if national legislation permits, to make questions on these topics voluntary. Alternatively (and ideally, in order to make it possible to distinguish simple non-response from deliberate non-provision of information) questions can be mandatory: but in this case it is imperative that they include a “prefer not to answer” option so that individuals may withhold the personal information. Answers to such questions should never be imputed where they are left unanswered or the “prefer not to answer” option is indicated.

1336. Countries may opt to implement special monitoring mechanisms related to the collection of data on ethnocultural characteristics, to guarantee free declaration, faithful enumeration (in cases where enumerators are used) and data protection.

1337. In drafting and testing census questions, defining concepts, developing classifications and designing statistical outputs, consultations should be undertaken with representatives of ethnic, language and religious communities. Engaging with such groups to foster understanding of the procedures, reasons for, and conduct of the census and the collection of information on these topics as part of it, can promote transparency, inclusion, equality of treatment and willing participation in the census. Thus, careful attention to this aspect can have positive spillover effects for the census exercise as a whole.

1338. Countries with long-established multi-ethnic communities, recently arrived immigrant populations and/or indigenous populations may wish to collect information on the ethnic composition of the population or of certain sub-groups of the population. This information is relevant for the understanding of the cultural diversity of the population, the position of ethnic groups, including indigenous groups, in society, and the definition and monitoring of equality and anti-discrimination policies. Ethnicity may provide a more accurate picture of the nature of immigrant or immigrant-descendant populations than can be obtained from information on country of birth or country of parents’ birth alone, which conceals ethnic diversity among immigrants from a given country; fails to reflect the ethnic diversity among the descendants of immigrants beyond the second generation; and masks non-immigrant indigenous ethnic diversity.

1339. Ethnicity is a concept based on a perceived affiliation with a group arising from shared historical and territorial origins (e.g. regional, national, tribal), as well as on cultural characteristics, such as language, religion and/or customs and ways of life.

1340. There is no internationally agreed definition or classification of ethnicity for statistical purposes. Countries may have their own definitions and classifications of ethnicity for statistical purposes, or these may be derived from national administrative or legal imperatives. The concepts employed and the intended uses of the information vary widely among countries. Countries should document the basic criteria and classification procedures for ethnicity and should inform data users about the concepts on which they are based.

1341. The concept of ethnicity should not be confounded with country of citizenship or country of birth. In languages (including English) where the term for “nationality” is commonly understood to be largely or wholly synonymous with citizenship, this term should not be used interchangeably with ethnicity, to avoid causing confusion among either respondents or data users.

1342. Many factors can affect the reporting of ethnicity in a census, including: respondents’ understanding or views about ethnicity; awareness of their family background; the number of generations that have lived in a country or the length of time since immigration; political sensitivities or social factors associating positive or negative connotations with certain ethnic identities. Since ethnic identity is not a fixed characteristic but a dynamic one influenced by many social forces, and since the range of ethnicities identified by individuals and by society is equally dynamic, classifications of ethnicity may need to change between censuses in order to reflect society at a point in time.

1343. Since ethnicity is subjective and based on self-identification, these factors do not necessarily render the collected data “low quality” or incomparable; but they do mean that trends over time in the ethnic composition of a population will be a result of interactions among these many factors. For example, growth over time in a given ethnic group may be a result not only of new arrivals who identify with that ethnicity, but also of a tendency for people to identify with that ethnicity who previously identified differently.

1344. Ethnic identity can be measured using a variety of concepts, including ethnic ancestry or origin, ethnic group, cultural origin, nationality, minority status, tribe, language, religion or various combinations of these concepts. For census purposes, affiliation with certain ethnic group should ideally be distinguished from affiliation with a language and/or religious group, although there may be considerable overlap in some cases.

1345. In some countries, ethnicity is understood (whether culturally or within official administrative systems) to be related to physical or racial characteristics of the population. Data on physical or racial characteristics may be used to identify visible minorities or racialized groups.

1346. Countries may endeavour to collect information on the ethnicity of respondents’ parents and grandparents, to gain a deeper understanding of the origins of the population and of integration processes. However, the collection of information on persons other than the data subject and other household members via a census is not recommended.

1347. The method and the wording of the question(s) used to collect ethnicity data can influence how respondents report their ethnic identity. In addition to any pre-coded response options, census questions should, where possible, provide the facility for write-in (open) responses, as this enables free self-declaration. It should able possible to report mixed ethnicities (more than one ethnic affiliation or a combination of ethnic affiliations), as well as to indicate “none” , “not declared” or “prefer not to answer”, either in the write-in facility or as a distinct option. Some countries have found that such a separate option increases non-response and impacts data quality. This possibility should be considered during testing and weighed against the desire to give respondents the possibility to avoid answering.

1348. Instructions for completing the census should explain how respondents are expected to report the ethnicity of children whose parents are of different ethnicities.

1349. In censuses using enumerators to collect information from respondents, countries should consider very carefully the existence of any ethnic tensions or sensitivities in the national context. Instructions for enumerators should emphasize that ethnicity must be freely reported by the respondents and not assumed by the enumerators. It may be necessary to introduce specific quality control procedures to ensure that this instruction is followed.

1350. When producing classifications of ethnic groups for census outputs, it is important to consider the risk of inadvertent disclosure for small groups and/or small geographic areas. While disclosure must be avoided, it is also important not to set a release threshold so high that ethnic minorities become masked in census outputs.

1351. Religion is generally regarded as a set of beliefs and practices, usually involving acknowledgment of a divine or higher being, power or principle, by which people order the conduct of their lives both practically and in a moral sense. For the purposes of collecting information in a census, this concept may be defined as either:

(a) a religious or spiritual belief or faith, regardless of whether or not this belief or faith is represented by an organized group or body; or

(b) an affiliation with, or membership of, an organized group or body having specific religious or spiritual tenets and/or practices.

1352. Countries that are traditionally multi-denominational or that have significant immigrant populations with different religions may wish to collect data on religion, either in addition to, or instead of, data on ethnicity. The decision to collect and disseminate information on religion in a national census depends on several considerations and national circumstances, including the national needs and expected uses for such information and the sensitivity, legal obligation or prohibition, or social acceptability of using the census as the source for such information.

1353. Depending on the specific needs and circumstances in a country, data could be collected on any of the following aspects of religion:

(a)formal membership of a church, temple or a religious community

(b)identification with a certain religion, religious community or denomination

(c)religious belief, whether practiced or not

(d)religion in which a person was brought up

(e)religious attendance or observance.

1354. The sensitive nature of census questions on religion necessitates special care to reassure the public of the existence of appropriate data protection and disclosure control measures. The public should be informed of the needs and uses for such information and the concepts and definitions used, both in collection instruments and in dissemination products.

1355. Where information on religion is gathered directly from respondents it should be based on the free self-declaration of respondents. Some countries may set a minimum age threshold for the collection and dissemination of information on religion for children, as determined by national circumstances and user needs. If young children are included, clear instruction should be given as to how their religious identity or affiliation should be determined.

1356. It should be made very clear if response is voluntary. In some countries, religion questions must be voluntary by law. As with ethnicity (see paragraph ‎1347), it is preferable to include the facility for write-in (open) responses and to give respondents the option to declare “no response” or “prefer not to answer”, rather than making the question voluntary. As in the case of ethnicity, the risk of increasing non-response by including this as a separate option should be considered and tested. Religion questions should also offer a “no religion” option which is distinct from the non-response option, allowing respondents to indicate that they do not have any religion, and thus permitting the separate identification of the non-religious population from those who chose not to respond.

1357. There is no internationally-agreed definition or classification of religion for statistical purposes. A classification should be comprehensive, exhaustive, and should be developed to reflect as fully as possible the diversity of the population in question. It should include groups of religions, individual religions, and subsets of religions such as religious denominations, administrative and organizational groupings, and independent religious groups, as well as belief systems typically not considered as conventional religions. If there is a significant and diverse non-religious population, countries may develop a classification that can identify separate groups within this population, such as Agnostic, Atheist and Humanist. When selecting labels for categories within a classification, countries should test the social acceptability of the labels chosen. For example, in some countries it has been found that terminology based on the identity rather than the name of the religion is preferred (e.g. “Muslim”, “Jewish” and “Christian”, rather than “Islam”, “Judaism” and “Christianity”).

1358. Countries may wish to collect data on languages that are currently spoken, understood and/or written, as an additional means of measuring cultural identity and integration. This may be especially relevant in traditionally multilingual countries or those with significant immigrant populations, but even countries less characterized by these features may wish to produce such data in the context of the growing importance of knowledge economies and the globalization of trade.

1359. In countries with more than one official language, it may be necessary to produce data on the use of the official languages to meet legislative or policy requirements.

1360. Depending on information needs, for countries wishing to include this topic in the census it is recommended to collect or produce data on one or more of the following concepts:

(a) First languages spoken or learned at home in early childhood

(b) Knowledge of language(s), defined as the ability to speak and/or write (and/or otherwise communicate) a specified language (e.g. the national official language(s) or significant minority languages), to a pre-defined level

(c) Usual language(s), defined as the language(s) most often currently spoken

i. at home (for all persons)

ii. at work (for employed persons)

(d) Main language(s), defined as the language(s) which the person best commands.

1361. Concepts and definitions used should account for the use of languages that are conveyed and/or understood by means other than conventional reading and writing, such as signed languages and Braille.

1362. Data collection instruments should permit the reporting of multiple languages, and instructions to respondents should make this clear. If relevant to national circumstances and if feasible within the data collection method, some degree of ranking may be used where the knowledge or use of multiple languages is reported. This enables the census to capture information on respondents who have acquired multiple languages, whether during childhood or subsequently, thus providing a fuller picture of the linguistic diversity of a country’s population.

1363. The collection and dissemination challenges of permitting multiple responses in language questions (as is also true for ethnicity and religion) may be significant, and special care must be taken to protect confidentiality in census outputs.

1364. As with ethnicity and religion, the population of many language groups can be small. It is therefore recommended that countries should include a facility to provide a write-in (open) response to any question on knowledge or use of languages. For literacy questions, standardized measurement tools should be used (see also section ‎18.6).

1365. As in the cases of ethnicity and religion, concepts and definitions used should be clearly explained, both to respondents during data collection and to data users in dissemination products. Census instructions should make clear to users of sign language that they should include these in their responses.

1366. As for religion, any classification used for outputs should be as comprehensive and exhaustive as possible to reflect fully the national circumstances (within the constraints of disclosure control rules), and should be tested for social acceptability regarding both the completeness of the classification and the choice of labels applied to categories. Special testing may be required to ensure that the design of questions and the designation of categories in outputs are acceptable to relevant population groups in the case of sign language or any language that may be seen as sensitive (such as languages used by vulnerable minorities).