947. This chapter provides recommended definitions and classifications for the following three themes: (1) household characteristics, (2) family characteristics, and (3) household and family characteristics of persons. Methodological considerations relevant to each theme are also described.

948. Many phenomena can be examined equally from the perspective of the private household, the family nucleus, or the individual. For instance, statistics related to one-parent families could be examined from the perspective of the household (e.g., number of households composed of a one-parent family), the family (e.g., number of one-parent families), or the person (e.g., number of persons/children/parents in one-parent families). Each perspective may yield substantively different results owing to the different numerators and denominators involved. NSOs should therefore always make clear the analytical perspective when disseminating products related to household characteristics, family characteristics or household and family characteristics of persons. NSOs should also make it clear that any data or classifications relating to the characteristics of families are restricted to families in private households, and the data and classifications related to the household and family characteristics of persons are restricted to the population living in private households.

949. The recommendations in this chapter are rooted in the concepts of institutional households and private households. It is recommended that the place of usual residence be used to determine whether a person resides in an institutional household, a private household, or neither. If only legal residence is available (for example when using register-based enumeration), then that information can be used if it is judged to reflect the usual residence situation of the population with acceptable accuracy.

950. A group of persons for whom housing and often daily necessities are being provided by an institution (for the definition of an institution see paragraph ‎1065‎(b)).

951. There may be differences between countries in the ways in which the boundary between the population living in private households, or in institutional households, or in non-institutional collective living quarters is drawn. The definitions used should therefore be explained clearly in the census reports and metadata, and attention should be drawn to any differences between national practice and these recommendations.

952. A person whose place of usual residence is at an institution. This includes people who have no other place of usual residence and who are residing at an institution on the census reference date, provided they are in-scope for census enumeration.

953. Prior to the census enumeration, it is recommended that countries carry out an institutional household validation exercise. This validation exercise or other form of “frame maintenance” should be based on a database of all institutions that is updated in each pre-census cycle, utilizing administrative data where possible. Each known institution should be contacted to confirm its status and to obtain a correspondent’s contact information. This inquiry could take the form of a brief survey questionnaire to identify, among other things, the nature and functions of the collective living quarters, the presence of private households attached to or within the institution (section ‎16.2.1.2.2), the type of populations who may reside there (i.e., adults only, adults and/or children, or children only) and whether certain key services are offered to members (the provision of care, food, etc.) For countries following a direct enumeration approach, this communication should also include information about the forthcoming census enumeration and the method in which it will occur.

954. Any person living in a private household within or attached to the institution’s collective living quarters should be excluded from the population living in an institutional household (see section ‎16.2.1.1.2). For example, employees of the institution who live alone or with their family at the institution (or in housing units located beside but separate from the collective living quarters of the institution) should be treated as members of private households.

955. The following categories are recommended for classifying institutional households:

Total – Institutional households

(1.0)Residential care facilities for adults and/or children (e.g., hospitals, hospices, palliative care units, residential units for people with (mental, physical or sensory) disabilities, rehabilitation facilities, reintegrative housing facilities, other long-term care facilities, such as homes for the elderly and nursing homes)

(2.0)Correctional and custodial facilities (including those for juveniles)

(3.0)Residences for students

(4.0)Military bases and government vessels

(5.0)Shelters or refuges for adults and/or children (e.g., shelters for unhoused persons, asylum seekers)

(6.0)Religious establishments

956. Given the large variation across countries in the prevalence and nature of institutional households, this classification should be considered a guideline only regarding the most common institutional household types. The examples in parentheses should be considered non-exhaustive.

957. Countries can opt to further distinguish types 1.0 to 6.0 where relevant for the country-specific setting. When possible, countries should distinguish residential care facilities and other institutional establishments which provide accommodation and services exclusively to children (e.g., aged 0 to 17 years).

958. The definition of a private household is different for countries that follow the housekeeping concept versus those that follow the household-dwelling concept. The definition used will have an important impact on the number and type of private households. Countries should therefore clearly communicate the chosen definition in disseminated products. See section ‎16.2.2.2.1 for more details.

959. Under the housekeeping approach, a private household is defined as a person or group of persons who occupy a whole or part of a private housing unit as their place of usual residence. Persons form a separate private household within a given housing unit if they live in a separate room (or rooms) which does not join with any of the other occupants of the housing unit. In multi-person private households, the persons may be related to one another, unrelated to one another, or some combination of the two.

960. When following this concept, countries should distinguish between boarders and lodgers where relevant. Boarders take meals with the rest of the housing unit and generally share the household facilities. Boarders should therefore be members of the same private household as the other occupant(s) of the household unit. Lodgers, however, rent part of the housing unit for their exclusive use. Lodgers should therefore belong to a different private household than the other members of the housing unit.

961. As illustrated in Figure 4, under the housekeeping approach, the number of private households is usually larger than the number of private occupied housing units.

962. Under the household-dwelling approach, a private household is defined as a person or group of persons who occupy a whole private housing unit as their place of usual residence. In multi-person private households, the persons may be related to one another, unrelated to one another, or some combination of the two.

963. In the household-dwelling approach, the number of private households is equal to the number of private occupied housing units.

964. A household that includes at least three separate generations of the same family (that is, persons linked by biological, spouse/partner or adoptive relationships). Most commonly, this includes one or more children, their parent(s) and their grandparent(s).

965. A private household that includes at least one grandparent and one or more of their grandchildren, with no parent of the grandchild(ren) present.

966. Note: countries that identify skip-generation households will need to decide whether these situations are considered to form either (a) a family nucleus, or (b) an extended family but not a family nucleus. This choice must be carefully considered and communicated. See section ‎16.2.2.3 for more details.

967. Results of the 2023 country practices survey showed a near-equal division of countries that conceptualize households based on the housekeeping approach (section ‎16.2.2.1.2) versus the household-dwelling approach (section ‎16.2.2.1.3).

968. In recent decades, many countries have seen a growing prevalence of living arrangements such as living with extended family members, multiple families sharing a housing unit, or a family sharing their living space with one or more unrelated persons. Given these trends, the chosen definition of private household is important, as the number of private households will usually be larger under the housekeeping concept in comparison with the household-dwelling concept, particularly the number of one-person households. The choice of private household definition will also impact the relevance and feasibility of certain classifications, as discussed later in this chapter. As a result, countries should clearly indicate in their disseminated products whether they use the housekeeping approach or the household-dwelling approach.

969. Results of the 2023 survey of country practices indicated a near-equal division in terms of the number of countries who use a register-based census methodology, a census methodology based on direct enumeration, or a methodology combining direct enumeration, register and/or administrative data i.e., a “combined” approach. Like the choice of housekeeping versus household-dwelling approach, the overall census methodology used will impact the conceptualization of private households. For instance, countries utilizing a population register may not be able to determine all or any extended-family relationships, precluding the identification of certain types of households.

970. Some countries identify skip-generation households (section ‎16.2.2.1.5). If identified, countries should make clear the following in their disseminated products:

(a) Whether skip-generation relationships are considered to form a family nucleus or not; and

(b) If skip-generation relationships are considered to constitute a family nucleus, this exception to the standard definition of a family nucleus must be clearly communicated. The consequences of this broadened conceptualization of a family nucleus on the classifications of household type, family nucleus type and various family and household characteristics of persons must also be carefully considered and communicated. Otherwise, there may be conceptual discrepancies that arise across various inter-related classifications. For instance, if person A is a grandparent in a skip-generation family nucleus, it must then be determined whether that grandparent is then classified as having a family status of “grandparent in a skip-generation family nucleus” or alternatively as a “parent in a family nucleus”. Under the latter approach, it is imperative to clarify in notes that some persons in the “parent” category are grandparents in skip-generation families. The relative prevalence of skip-generation households in each country setting should guide this decision-making process.

971. The following classification of private households⁹⁵ by size is recommended:

Total – private households

(1.0)1 person

(2.0)2 persons

(3.0)3 persons

(4.0)4 persons

(5.0)5 persons

(6.0)6 or more persons.

972. The following classification of private households⁹⁶ by type is recommended:

Total – private households

(1.0)One-family-nucleus households without additional persons

(1.1)Couple without children

(1.2)Couple with children

(1.3)One-parent family

(2.0)Other family households

(2.1)Three-or-more-generation households

(2.2)Multiple-family households (excluding three-or-more-generation households)

(2.3)One-family households with additional persons not in a family nucleus (excluding three-or-more-generation households)

(3.0)Non-family households

(3.1)One-person households

(3.2)Two-or-more-person non-family households

973. See paragraph ‎979 for the definition of family nucleus and section ‎16.4.1 for the definition of child in a family nucleus.

974. The distinction of categories 2.1, 2.2 and 2.3 is optional. Particularly for countries following a housekeeping approach, it may not be possible to distinguish all or any of these “other family households” which require the identification of various extended-family relationships.

975. In household categories 2.0, 2.1, 2.2 and 2.3 other unrelated or related persons may also be present in the household.

976. In household category 3.2, none of the members form a family nucleus.

977. Countries may choose to further distinguish:

(a) The type of couple in categories 1.1 and 1.2 (for instance, distinguishing married versus non-married couples, and the sex or gender of the persons in the couple);

(b) The sex or gender of the parent in category 1.3;

(c) The age of children in family nuclei, if there are pertinent age grouping distinctions in that specific country setting (e.g., couples without children, couples with at least one child aged 24 or younger, couples with all children aged 25 or older);

(d) The family structure in category 2.3, i.e., whether the family is composed of a couple with or without children or is a one-parent family.

978. Some countries using a register-based method may only be able to identify married couples. If this is the case, the household types 1.1 and 1.2 must be labelled accordingly i.e., “Married couple with children”, “Married couple without children.” See section ‎16.3.2.4 for additional guidance.

979. Family nucleus. Two or more persons living in the same private household who together form a couple (with or without children) or a one-parent family. Within a family nucleus, all inter-person relationships are limited to direct (first-degree) relationships, that is, between spouses or partners or between parent(s) and children.

980. Reconstituted family. A family nucleus composed of a couple with children in which at least one child is the biological or adopted child of only one spouse or partner in the couple.

981. Extended family. Two or more persons living in the same private household who are related through a biological, spouse/partner or adoptive relationship, even if the relationship is non-direct (for example, two cousins living together). All members of a given family nucleus (see paragraph ‎979) are also members of the same extended family.

982. Countries that use a register-based or combined census methodology are less likely to be able to produce information on reconstituted families and extended families. For this reason, the recommended classifications related to these family structures are optional.

983. Information should be collected for all individuals living in private households on their relationship with other members of the household. Data on this topic are needed for use in:

(a) Identifying private household types;

(b) Identifying family nuclei and family nucleus types within a given private household; and

(c) Identifying the household and family characteristics of persons within a given private household.

984. There are various methods of determining relationships within a private household, each offering their own benefits and challenges. The three main approaches used by countries are illustrated in Figure 5 and described below.

985. Reference-person approach: the relationship between one person - the “household reference person” - and each other member of the household is determined. Family and household types and individual family statuses are then derived from that information. According to the 2023 survey of country practices, most countries continue to use this traditional reference-person approach.

986. Determining which member of the household is the reference person requires careful consideration. In the past, the person considered to be the “head” of the household was generally used as the reference person, but this concept is no longer used in many countries. It has also sometimes been proposed that the oldest person or the person who contributes the most income to the household should be the reference person. However, given that the question's main purpose is to assign family status, family type and household type, both approaches have weaknesses. The automatic selection of the oldest person may be undesirable because in three-or-more-generation households, the broadest range of explicit kin relationships can be reported where the reference person is selected from the middle generation. Similarly, the selection of the person with the highest income may be a person who will not necessarily solicit the broadest range of explicit kin relationships.

987. For countries using a direct enumeration for their census methodology, it is recommended to provide clear instructions to the respondent to list the members of the household in a specific order:

(a) Begin with an adult (the reference person), followed by their spouse or partner (if applicable) and then any children of that person (if applicable);

(b) Next, add any other household members.

988. Despite providing clear instructions to respondents about how to order members of the household on the initial roster listing, inevitably these instructions will not always be followed as intended. To help mitigate this challenge, response options in terms of relationships to the reference person should be exhaustive for all (other) household members, and not conditional on their listing order.

989. When the household's reference person is chosen carefully, this approach gives accurate information for most household and family types. In certain cases, however, such as in multiple-family households, this method may not always provide sufficient information; as a result, greater levels of imputation and inference are needed to determine each household member’s family characteristics.

990. Full relationship matrix approach: all possible inter-person relationships within a given private household are determined.

991. Used by a growing number of countries, the full relationship matrix approach provides the most complete detail on living arrangements within the household, particularly for large households involving a mix of related and unrelated people. This approach also makes possible in theory the identification of relationships that are traditionally not captured by the reference-person approach, such as siblings living in the same household together as foster children, or various relationships existing within “other family households”.

992. Conversely, the full relationship matrix approach also presents several disadvantages. For countries whose census is based on direct enumeration, the additional information to be collected makes the questionnaire more time-consuming and burdensome for respondents to complete - especially for larger households - as well as literally lengthier in the case of paper-based questionnaires, increasing printing costs. In comparison with the reference-person approach, the full relationship matrix approach is also typically much more expensive from an operational perspective, requiring higher resources in terms of coding and processing than the reference-person approach.

993. These challenges can be mitigated somewhat by adopting a partial relationship matrix approach. In contrast to the full relationship matrix approach, under a partial relationship matrix approach, a complete relationship matrix is only collected for a subset of all household members (the first 5 persons listed in the household roster, for instance). Additional household members can then be asked, alternatively: only about their relationship to a household reference person; only about their relationship to their “nearest neighbour” in the household roster (e.g., person 6’s relationship to person 5), or a combination of these two options.

994. While a relationship matrix approach is more likely to be adopted by countries using direct enumeration for their census methodology, it is sometimes possible for countries using a register or combined methodology to take this approach to relationship identification, provided information about two given household members can be linked to one another on their respective register data.

995. In summary, countries contemplating the adoption of a relationship matrix approach should first execute thorough feasibility analyses of adopting this method, both in terms of public acceptability (via qualitative and quantitative testing) and by conducting a cost-benefit evaluation of the additional operational and resource requirements contrasted with the added nuance and richness of information. In some cases, the new processing requirements may be net neutral if for instance, by having complete household relationships it reduces the need for complex imputation processes that previously filled in the gaps. Where feasible, the full relationship matrix method is the recommended approach to identify household relationships since it provides the most comprehensive information.

996. Along with societal changes over recent decades, the intersection of family, sex and gender in censuses is also changing. According to the 2023 survey of country practices, about half of surveyed countries made a distinction in their last census between opposite-sex and same-sex couples. In recent years, some CES countries have begun to distinguish sex from gender in their census, and to classify couples according to the gender of each member of the couple. Other countries have plans to collect gender and sexual orientation in future censuses. At the same time, for some countries, these types of distinctions are not planned to be collected or disseminated. The terminologies, definitions and classifications used by statistical agencies to measure these concepts are emerging and constantly evolving. See section ‎15.2 for more information and guidance.

997. In contrast to the previous edition of these Recommendations, the present chapter does not make any sex or gender-based distinctions in its definitions or classifications related to household and family characteristics. Countries can nevertheless opt to make these additional distinctions when possible.

998. Countries vary in terms of the information collected and disseminated about couple families. These variations relate to the specific legal or fiscal implications of various types of couples in each country setting as well as the country’s (in)ability to identify certain types of couples in their census. For instance, it may be impossible to identify non-married or otherwise non-registered couples if countries are using a register-based census methodology; for others, only partial information may be available (e.g., for those unregistered or unmarried couples who share a common child). In the present chapter, definitions and classifications related to couples do not make distinctions based on couple type; that is, whether the couple is married, non-married in some type of legally or formally-recognized partnership, or in another form of non-married co-residing relationship. Countries can nevertheless opt to make these types of distinctions where possible. It is recommended that, when these distinctions are made, the various types of couples are clearly defined, particularly when there are legal implications attached.

999. The phenomenon of couples living apart, i.e., spouses or partners who do not share a place of usual residence, is of growing interest, and countries are encouraged to evaluate the feasibility of collecting and disseminating information about this phenomenon via their censuses if not collected elsewhere. If collected, this phenomenon should be clearly distinguished from co-residing couples. Otherwise, countries should make clear in their disseminated products that information about couples and individual membership in a couple is restricted to persons who share a place of usual residence in a private household with their spouse or partner.

1000. Similarly, countries should also make clear the distinction between a person’s marital status versus their family characteristics (i.e., whether they are part of a couple family nucleus within a private household). Generally, marital status is collected for the total population (including those living in institutional households) while family characteristics are typically only collected for the population living in private households. There are also situations where a person may be married, living in a private household, but not living with their spouse. This can occur when, for example: the person’s spouse or partner is residing in a long-term care institution or other institution; during a pre-divorce separation period; or when economic or housing barriers prevent spouses or partners from establishing a new residence together, among other reasons. In short, the number of married persons in the total population will not necessarily equal the total number of persons in married couples living in private households.

1001. In cases where a private household includes three-or-more generations of the same family, family nucleus status should be prioritized for the two youngest generations where applicable. provides examples of how family nucleus status should be determined (Table 5).

1002. Countries should include guidance in reference documentation on the limitations of the “address-based” approach used in censuses to identify families. Specifically, countries should emphasize that the census concept of a family is based around the place of usual residence among persons in private households, which may not always capture a comprehensive portrait of the family structure and living arrangements of all persons. This is particularly true for persons who split their time living in two or more households, for example, children with shared parenting arrangements; students who sometimes live at their parental home and other times at a student dormitory; spouses or partners who live together part of the year and separate for part of the year for work or other reasons. NSOs should clearly document how the place of usual residence is determined in these situations.

1003. The following classification of family nuclei⁹⁸ by type is recommended:

Total – family nuclei in private households

(1.0) Couple family

(1.1)Without children

(1.2)With child(ren)

(2.0) One-parent family

1004. Countries can opt to make any of the following additional sub-categories:

(a) Of types 1.0, 1.1 and 1.2 based on the type of couple, including sex or gender-based distinctions;

(b) Of types 1.2 and 2.0 based on age groupings of the youngest child;

(c) Of type 1.2 based on whether or not the couple family is a reconstituted family (see the classification Reconstituted family type);

(d) Of type 2.0 based on the sex or gender of the parent.

1005. The following classification of reconstituted families⁹⁹ by type is recommended:

Total – Reconstituted families

(1.0)With no common children

(1.1)Only one spouse/partner has at least one biological/adopted child

(1.2)Each spouse/partner has at least one biological/adopted child who is not the biological/adopted child of the other spouse/partner

(2.0)With at least one common child

(2.1)The spouses/partners have at least one common child and only one spouse/partner has at least one biological/adopted child who is not the biological/adopted child of the other spouse/partner

(2.2)The spouses/partners have at least one common child and each spouse/partner has at least one biological/adopted child who is not the biological/adopted child of the other spouse/partner

1006. In most cases, countries will only be able to identify reconstituted families if they use a relationship matrix approach to identify relationships within the household.

1007. Countries can opt to make additional distinctions of reconstituted families based on the sex or gender of each member of the couple and the type of couple.

1008. It can be difficult to communicate the possible variations in the structure of reconstituted families (i.e., the differences between categories 1.1, 1.2, 2.1 and 2.2); visual aids such as Figure 6 below have been found to be helpful in aiding comprehension.

1009. Countries should follow international standards with respect to age-based restrictions for household and family characteristics of persons. Specifically, only persons aged 15 and older can live alone, be part of a couple, or be a parent. Persons aged 14 and younger should not be the “reference person” when deriving household relationships; there should be at least one person aged 15 and older in every private household. Countries should ensure their editing and imputation processes consider these age-based standards when deriving household and family characteristics of persons.

1010. A person who lives in the same private household as at least one of their biological or adoptive parents and who does not have a spouse, partner or child(ren) of their own living in the same household. The child in a family nucleus may be of any age or marital status.

1011. Foster children are excluded from the family nucleus and the extended family of the foster parent(s) and should be considered as living with non-relatives.

1012. A child who alternates between two or more households within the country (for instance following their parents’ separation or divorce) should consider the household where they spend the majority of the time, or their registered place of residence, as their place of usual residence. Where an equal amount of time is spent in two or more separate households, the place of usual residence should be the one where the child is staying at the census reference time. (See paragraph ‎781‎(c))

1013. As a result of their overall census methodology, not all countries will be able to measure certain classifications related to the household and family characteristics of persons. For instance, countries that use a register-based (as opposed to direct enumeration) census methodology may have more difficulty identifying whether an individual is living with members of their extended family. Additionally, countries using a reference-person approach (as opposed to relationship matrix approach) to identify relationships within the household may not be able to capture all inter-personal relationship details within the household, particularly in multiple-family households.

1014. Along with the phenomenon of grandchildren living with grandparents, most countries have seen an increasing number of older (adult) children living with their parent(s). Indeed, if considered as a form of multi-generational living, the phenomenon of a two-adult-generation household is likely much more prevalent than a three-or-more-generation household”, as indicated in a recent study of the United States¹⁰⁰. Adult children may live with their parents for a variety of reasons, including: to manage housing costs or in response to housing shortages; to support and care for their aged parent; to receive informal childcare assistance from their child’s grandparent; for companionship; for cultural normative reasons.

1015. The classification “Presence of person’s parent(s) in the household” is recommended for examining this living arrangement.

1016. Some member countries have witnessed growth over recent decades in the phenomenon of three or more generations of a family sharing a place of usual residence. The flows of care and support within these households are not necessarily unidirectional and can depend on the age and individual family situation of each respective generation. Less commonly, grandchildren and grandparents may live together without the middle generation present, in a skip-generation household. The prevalence and therefore the relevance of these living arrangements vary across and within countries.

1017. Countries interested in disseminating information on these phenomena should do this not only from the perspective of the household but also from the perspective of the person - specifically, from the perspective of young children (for example, those aged 14 and under). When examined from the point of view of young children, the rates of prevalence tend to be much higher and arguably more meaningful than those obtained from a household lens. The classification “Presence of person’s grandparent(s) in the household” is recommended for this purpose.

1018. The following classification of the population by household status is recommended.

(1.0) Person in a private household

(2.0) Person not in a private household

(2.1) Person in institutional household

(2.2) Other

1019. The classification Private household type can be transferred to the person level as follows.

Total – persons in private households

(1.0)In one-family-nucleus households without additional persons

(1.1)In a couple-without-children household

(1.2)In a couple-with-children household

(1.3)In a one-parent-family household

(2.0)In other family households

(2.1)In a three-or-more-generation household

(2.2)In a multiple-family household (excluding three-or-more-generation households)

(2.3)In a one-family-nucleus household with additional persons not in a family nucleus (excluding three-or-more-generation households)

(3.0)In non-family households

(3.1)In a one-person household

(3.2)In a two-or-more-person non-family household

1020. The following classification of the population by family nucleus status and household living arrangement is recommended:

Total – persons in private households

(1.0)Person is in a family nucleus

(1.1)Spouse or partner in a couple

(1.1.1)Without children

(1.1.2)With child(ren)

(1.2)Parent in a one-parent family

(1.3)Child

(2.0)Person is not in a family nucleus

(2.1)Living with others

(2.1.1)Living with other relatives (non-relatives may also be present)

(2.1.2)Living with non-relatives only (includes foster children)

(2.2)Living alone

1021. The term “other relatives” refers to other persons to whom the person is indirectly related, that is, they are not part of the same family nucleus, but they are nonetheless linked through an indirect biological, spouse/partner or adoptive relationship. Examples include cousins, nieces and nephews, aunts and uncles.

1022. For the most complete portrait of a person’s household and family characteristics, it is recommended to cross this classification with the Private household type of person classification.

1023. The following classification of the population by presence of parent(s) in the household is recommended:

Total – persons in private households

(1.0)Living with at least one parent

(2.0)Not living with a parent

1024. Countries utilizing a register-based census methodology may not be able to identify these living arrangement details.

1025. Countries should include the age group and family nucleus status of the person as an additional variable in the disseminated table product to make the information in this classification substantively meaningful.

1026. The following classification of the population by presence of grandparent(s) in the household is recommended:

Total – persons in private households

(1.0)Living with at least one grandparent

(1.1)And two parents

(1.2)And one parent

(1.3)And without parents

(2.0)Living without grandparents

(2.1)With two parents

(2.2)With one parent

(2.3)And without parents

1027. Countries utilizing a register-based census methodology may not be able to identify grandchild-grandparent relationships.

1028. Countries should include the age group and family nucleus status of the person as an additional variable in the disseminated table product to make the information in this classification substantively meaningful.

1029. The following classification of the population by extended family status is recommended:

Total – Persons in private households

(1.0)Person is in an extended family

(1.1)In a family nucleus

(1.2)Not in a family nucleus

(2.0)Person is not in an extended family

1030. This classification can be used to derive some categories of the classification of the family nucleus status and household living arrangement of person (section ‎16.4.3.3). For instance, persons whose extended family status is sub-type 2.0 (not in an extended family) and who live in a multi-person household would have a family status and household living arrangement of “living with non-relatives only.” Persons whose extended family status is sub-type 1.2 (in an extended family, not in a family nucleus) would have a family status and household living arrangement of “living with other relatives (non-relatives may also be present).”

1031. The following topics do not relate specifically to relationships within households. They instead refer to housing-related topics which are examined from the perspective of private households.

1032. The following classification of the private households by sole or shared occupancy of the housing unit is recommended:

(1.0)Private households which have sole occupancy of a housing unit

(2.0)Private households which share a housing unit with one or more other private households

1033. This classification applies to private households. It can only be used by countries that use the housekeeping approach to define private households. For countries that use the household-dwelling approach, all private household will fall under Category 1.0.

1034. Category 2.0 may be sub-divided to distinguish households sharing a housing unit with one, two, or three or more other households.

1035. The following classification of the private households by tenure status of the housing unit is recommended:

(1.0)Households of which at least one member is the owner of the housing unit

(2.0)Households of which no member is the owner of all or part of the housing unit

(2.1)Households of which a member is a main tenant of all or part of the housing unit

(2.2)Households of which a member is a sub-tenant of an owner occupier or main tenant

(3.0)Households occupying all or part of a housing unit under other form of tenure

1036. This classification applies to private households. It is recommended for use at the one-digit level and optional at the two-digit level.

1037. In view of the diversity of legal arrangements internationally, countries should describe fully in their census report or relevant metadata the coverage of each of the categories in the above classification. These descriptions should specify, where applicable, the treatment of households which:

(a) Live in housing units as members of different types of housing cooperatives;

(b) Live in housing units rented from an employer under the terms of the contract of employment of one of the household members; and

(c) Live in housing units provided free of charge by an employer of one of the household members or by some other person or body.

1038. Some countries may wish to extend the recommended classification to distinguish these or other groups of households that are of interest for national purposes. Households that are in the process of paying off a mortgage on the housing unit in which they live or are purchasing their housing unit over time under other financial arrangements, should be classed as (1.0) in the classification.

1039. Rent is the amount to be paid in respect of a specified period for the space occupied by a household including, in some cases, local rates and ground rent. Payments for the use of furniture, heating, electricity, gas and water and for the provision of special services like washing and cooking should be excluded.

1040. With regard to the costs of heating and hot water, the practices differ in different countries. In some countries the heating and hot water are normally included in the rent, in other countries they are not, while still in some other countries both practices can exist in parallel. It is important, therefore, that countries which include this topic in their census clearly define whether the heating and hot water are included in the rent or not. The recommendation would be to exclude heating and hot water expenditures from the rent. The ideal solution would be to produce the rent information separately without heating and hot water expenditures and to report the expenditures for heating, hot water (and electricity if applicable) separately.

1041. Nominal rent paid may not correctly reflect the real rates. For instance, an individual housing allowance determined on the basis of a means test and paid by housing authorities directly to the property owner should be included in the rent; and if a public sector property owner on the basis of a means test charges a rebated rent, the full rent should be recorded. It may also be possible to collect information on, for example, whether the tenant is a relative or an employee of the property owner, or whether he performs any function or office as part of his rent, in order to appraise the actual rent paid.

1042. If this topic is included in the census, it may be desirable (and less sensitive) to obtain information on the range within which the rent paid falls rather than on the exact amount paid.

1043. In addition to the amount of rent paid by renting households, it may be useful to collect information on the housing costs of each household. Such costs could include, for example, information on monthly mortgage payments, the provision of utility services and local taxes if these are not available from the information collected from the housing topics.