Relationship to head of household

Edward Higgs

The decennial census in Great Britain from 1801 to 1831, were organised by John Rickman, a clerk of the Houses of Parliament. Rickman was interested in collecting information on the numbers of families and houses in the country, but he did not collect information on the internal structure of families. In 1841 the organisation of the census in Great Britain was taken over by the General Register Office (GRO) in London. From 1861 the census in Scotland was undertaken by a separate GRO in Edinburgh. The censuses from 1841 onwards were based on household schedules in which householders were to provide certain items of information about the inhabitants of their households. In the period up to and including the 1901 census the information so provided was copied into books by census enumerators for dispatch to London and Edinburgh. After that date, data abstraction was done directly from the householders' schedules (Higgs, 1989, 5–15).

The introduction of the household schedule in 1841 made it possible to study the structure within families for the first time. However, in 1841 there was no relationship column in the census schedules. Enumerators were merely told to "Set down one after the other those who have the same surname, beginning with the heads of the family, and put no others between them". From 1851 onwards there was always a 'Relationship to head of family' column in the household schedule. In 1851 this column carried the instruction, 'State whether wife, son, daughter or other relative, visitor or servant'. In 1861 'boarder, etc.' was added to the list, as was 'head' in 1871. The instruction remained similar thereafter. At various times certain abbreviations were suggested, 'daur.' for daughter, 'serv.' for servant, 'F.-in-Law' and 'M.-in-Law' for father-in-law and mother-in-law respectively, and so on.

The headship of a household was, of course, a social position and did not necessarily reflect biological descent. Thus, in one household an aged widow might be described as the head but in another case a son or daughter who had taken over running the affairs of the group might be so described. Occasionally two men or women sharing a set of rooms might both be put down as joint heads. Sometimes the usual head of the household was absent and this was denoted in the manuscript returns by the first person in the household being described as a wife, son, servant, or some other term.

This information was used to make some passing comments on family structure within the population in the Census reports, and to make some rough distinctions between single person families, nuclear families, and those containing boarders and lodgers, and also to calculate the number of children per family (e.g. Census of England and Wales, 1861, Vol. III. General report, 10–11). The Victorian census authorities believed that, 'The natural family is founded by marriage, and consists, in its complete state, of husband, wife and children' (Census of England and Wales, 1871, Vol. IV. General report, xx). They may have seen this information, therefore, as a way of understanding the dynamics of family formation that helped to explain the processes by which the population expanded, especially in the urban centres. Population density was seen, in turn, as determining mortality rates (Higgs, 1991). It was not until the 1951 census, however, that a serious attempt was made to analyze the composition of private households (Office of Population Censuses and Surveys and the General Register Office, 145–7).

The analysis of this data never seems to have been a very prominent part of the Census reports prior to the Second World War, and may have declined over time. This might reflect uncertainties over the quality of the data, and the recognition that it could not be taken as a substitute for a proper fertility survey (as was undertaken in 1911). Certainly, the confusion between lodger and boarder, and the ambiguities relating to visitors, have been noted by historians studying the manuscript returns. Such designations probably have to be accepted at face value but relationship data from districts with large numbers of boarders, lodgers and visitors, as in the case of coastal resorts, should be treated with caution (Higgs, 1989, 58–60). The column may, in part, have survived in the manuscript returns as a means of checking that the householders and enumerators had correctly defined the household.

Nineteenth-century definitions of kin relationships were different on occasion from those of today. Relationships by marriage appear to cause particular problems. A brother-in-law or son-in-law could be described as a 'brother' or 'son', whilst a 'daughter-in-law' might mean a step-daughter. Similarly, the offspring of married children resident in the household were sometimes called 'son' or 'daughter' rather than grandchild. In such cases the relationship of the children may have 'slipped', and refer to their parents rather than to the head-grandparent. Sometimes the presence of an unmarried daughter of child-bearing age, raises the suspicion that the infant 'sons' and 'daughters' of elderly parents might be illegitimate grandchildren. Such problems could be multiplied.

The term 'servant' could also cause confusion, especially since it could appear in the occupational column. Service in the nineteenth century was a legal relationship between master and servant, rather than a defined set of tasks — the term 'civil servant' still preserves some of this original meaning. Apprentices, shop workers, domestic servants, and living-in agricultural labourers could all be regarded as 'servants'. Since it was the nature of the social relationship that was important, the distinction between a paid employee and the poor country cousin who did housework in return for her keep must also have been a fine one. Similarly, the extent to which the relationship 'servant' and the occupational designation 'housekeeper' were used to conceal unconventional marital relationships is difficult to establish. People with domestic service occupations were also not necessarily servants in the households in which they lived; they could be lodgers and kin who were unemployed, or go out to work in other houses during the day.

Conventions have been established in order to assign relationships to the members of the 1841 households found in the manuscript returns (Armstrong, 229–30). These rules are very rough and ready, and caution should be used in any attempt to compare differing populations from the 1841 census, or to compare these results with later censuses. It is assumed that the first listed person is the head. To allocate persons of the same surname, the following rules are applied:

(a) the first listed woman within 15 years of the head's age is assumed to be his wife. Other women, provided that (from consideration of their ages) they are born when the head and wife (where applicable) were aged not below 15 and not over 50, are assumed to be daughters. Any other women are treated as relatives;

(b) other males bearing the same name are regarded as sons, provided that they too were born when the head and wife were aged not less than 15 and not more than 50. Other males are regarded as relatives.

These conventions become more difficult to apply, of course, where the first person in a household is a woman. Further rules cover the allocation of those not bearing the head's surname. Domestic servants are taken as being all those so described in the occupation column, unless they have the same name as the head (in which case they are taken to be children or relatives), or unless the head of the household is also by occupation described as a servant (whereupon they are regarded as lodgers). All those not covered by the above rules are placed in a residual category of lodgers.

REFERENCES

W. A. Armstrong, 'Social structure from the early census returns', in E. A. Wrigley, ed., An introduction to english historical demography (London, 1966).

Census of England and Wales, 1861, Vol. III. General Report BPP 1863 LIII. Pt. I.[View this document: General report . England and Wales. 1861]

Census of England and Wales, 1871, Vol. IV. General Report BPP 1873 LXXI Pt. II. [View this document: General report, England and Wales. 1871]

Edward Higgs, 'Diseases, febrile poisons, and statistics: the census as a medical survey', Social History of Medicine, 4 (1991), 465–78.

Edward Higgs, Making sense of the census. The manuscript returns for England and Wales, 1801–1901 (London, 1989).

Office of Population Censuses and Surveys and the General Register Office, Edinburgh, Guide to Census Reports, Great Britain 1801–1966 (London, 1977).