Selected Subjects: Usual Residence

Next Selection Previous Selection

4.14 Usual Residence

4.14.1 1931

In each Census since 1931 there has been a question on usual residence. The answer requested has been 'HERE' for persons enumerated at their usual residence, and, in other cases the usual address to be written in.

To clarify what was intended by 'usual address in particular cases, there have been supplementary notes which have varied from census to census. For details see chart on opposite page.

In the census statistics of Great Britain it has always been the practice to assign the population to the place of abode on census night. Although this practice may be regarded theoretically as casual, statistical analyses of populations based upon this assignment have been generally accepted as equivalent to distributions according to the home areas of the population. It was not believed that the number of people away from home would make any appreciable difference to the results. In order to ensure that the returns should include as little holiday movement as possible, censuses have normally been taken in early spring. In 1801 the census was taken in March, in 1811-1841 near the end of May, and from 1851-1911 at the end of March or beginning of April. It was reasonably assumed that occasional population movements were at a low level at this time of the year in contrast to the larger displacements expected in the summer holiday months.

The year 1921 was exceptional in this respect, for though the census had originally been planned to be taken in April it had at the last moment to be postponed on account of industrial disturbance and could not be taken until 19 June. Although this date was in advance of the peak of the summer holidays, the movement had begun and there was no doubt that the populations of the towns and areas usually affected by it would have been appreciably lower at the earlier date.

With this experience in mind and also in view of the fact that transport development was progressively increasing, it was felt in 1931 that the chance of a greater percentage of the population being absent from their homes at week-ends, even outside the holiday periods, might no longer be negligible and that a test was desirable to ascertain how far census analyses of local populations based on place of enumeration might be regarded as continuing to meet needs and problems more specifically asso¿ciated with residence. Accordingly the following question was included in the 1931 Census schedule:

Usual Residence

  1. For persons included in this Schedule who usually reside at this dwelling or establishment, write 'Here'.
  2. For those who have a more usual residence elsewhere, give full postal address of usual residence.
  3. For those who have no settled place of residence, write 'None'.

The term 'usual residence' had no formal or statutory meaning. In order to give guidance in cases where its meaning might have been misunderstood, the question was supplemented by the following instructions printed on the back of the schedule:

Usual Residence: Column C

Visitors. Persons described in column B as 'Visitor' cannot, of course, be usually resident at the premises to which the schedule relates and the reply 'Here' would be wrong in such cases. All persons enumerated while on a visit to either a private family or to an Hotel, Club, Boarding House or Lodging House should give the full postal address of their usual residence, viz, their home address. This applies also to the case of visitors from abroad, who should give their full home postal address in the country overseas from which they have come.

Hotels, Boarding Houses, etc - See above - 'Visitors'. But persons enumerated at a private or other Hotel, Boarding House or Residential Club at which they have a settled residence, and who have no more usual place of residence elsewhere, should regard the Hotel, Club, etc as their usual residence and write 'Here'. Persons who move from one Hotel to another and have no settled residence anywhere should write 'None'.

Resident Domestic Servants, Shop Assistants, etc who live in should regard their 'situation1 (the private family, Hotel or Boarding House, etc at which they are employed), or the living-in premises, as their usual residence. Such persons, if enumerated at the place of employ¿ment or living-in premises, should write 'Here1. If enumerated while staying elsewhere with friends or relatives, they should give the address of the place of employment or living-in premises as their usual residence. But 'daily' domestic servants should regard their own home as their usual residence and reply accordingly.

Scholars, Students, Undergraduates, etc of a residential school, college or university who happen to be enumerated at their home address should state in column C the name and address of the residential school or college in which they usually reside during term time. University students who reside in term time in lodgings in the university town should regard such lodgings as their usual place of residence.

It was not possible to discern accurately how far these instructions were followed, but the classified results showed evidence of inconsistencies. In straightforward cases of temporary 'visitors' staying with friends there was little reason to suppose that there was any objection to giving their own address. In some cases, however, the interpretation of 'usual residence' may have occasioned real doubt and, when the person responsible failed to read or understand the instructions, there may have been a divergence from the intention of the census question. In other cases there may have been reluctance as with resident domestics or shop assistants, to return the employer's house or living-in premises in preference to their own home. It was not felt that discrepancies which arose from any of these reasons seriously affected the statistical record except in one particular class of case - the resident scholars of educational institutions. These, in spite of the specific instructions, failed in large measure to return their school address as their usual residence. For the purpose of constructing sub¿sequent annual estimates of population, the census statistics had to be adjusted to remedy this defect in those areas where it was likely to have been significant.

Of the total numbers enumerated away from their homes, only those whose usual residence was outside the borough or county district in which they were enumerated were taken into account in the census analysis. This was, in itself, a considerable undertaking; to have extended the analysis to cover the shorter range and less important movements within the several local authority areas would have added considerably to the work.


The classified statistics were published in the General Tables and statistics for local areas were given in Table 6 of the County Parts .

Detailed statistics derived from this enquiry in Scotland were given in the City and County Parts of Volume I of the Reports; comment upon them, together with a summary table was published in Volume II.



For the 1951 Census, the question on usual residence (in Col C) was in the same form as 1931, when it was first put. The answers were used to establish, for each local authority area, the number of persons who were enumerated there but had a usual residence elsewhere and conversely, the number of persons usually resident in the area who were enumerated elsewhere in England and Wales. This information enabled comparison to be made between the enumerated population and the usually resident population, and for the enumerated population to be adjusted so as to provide a base for the annual estimates of local populations which relate to the resident population.

The one big problem in handling data on this subject was the allocation of an area code in every case in which the address of enumeration and of the usual place of residence were in different local authority areas. The action taken to secure this information was dictated by the needs of workplace movement, where area coding to a place of work outside the area of usual residence was necessary in a very much larger number of cases, and it was decided to revert to the 1921 procedure when the question on place of work was asked for the first time and a system of local coding had been used.

Accordingly enumerators were issued with supplies of specially printed postcards for each of the two subjects and were instructed to prepare a card for every address of usual residence or workplace, given by any person whom they enumerated, which was outside the local authority area in which they were acting (apart from workplace addresses outside England and Wales). No account was taken of people enumerated away from their homes but within the same local authority area.


The Census Officer for the area first extracted cards bearing usual residence addresses outside England and Wales which he sent direct to Census Headquarters, and on all remaining cards inserted his own local code. He handed the cards to the GPO who forwarded them to the appropriate Census Officers in whose areas the addresses of usual residence or workplace were situated. These officers in turn added their own code and the cards were finally despatched to Census Headquarters. The arrangements, involving approximately one million usual residence and seven million workplace cards, generally worked satisfactorily.

Production of tabulations

Unlike the tabulations for other subjects the actual postcards for usual residence were manually sorted and used to produce the tables. The cards were later merged with those for workplace and matched against the schedule entries. (For fuller details see General Report) .

The main table in the published volume shows for each local authority area both the enumerated population and the estimated resident population, ie the enumerated population adjusted in respect of persons enumerated away from their usual area of residence.

Completeness of coverage

Comparison of the census population with the mid-1950 Total Population Estimate shows an apparent deficiency in the census population of 134 thousand or three per thousand. But, bearing in mind that the migration element in the estimates must be regarded as having a wide margin of error, any deficiency in the census record is probably quite trivial.

There is no evidence to suggest that the question on usual residence was not answered correctly, except possibly some misunderstanding by householders when census figures are compared with locally supplied information of the normal population of various types of non-private establishment. It is thought to have arisen simply through lack of study of the requirements by:

  1. heads of households containing students of residential schools or colleges who had returned to their family home on vacation for a period covering census day, and who were probably regarded without question as normally settled in that home instead of (as should have been the case) at their term-time address, and
  2. some members of the armed forces on leave from their normal station who similarly regarded their home address as their usual place of residence.

In addition, some mental hospitals found it difficult to apply the 'six-months' rule under which if a patient was expected to be discharged within six months of census day he or she should have been assigned to the address of discharge. If this address was not known 'none' was the term to be used.

Other patients and inmates were regarded as usually resident in the hospital. This rule applied with slight variations to all types of institutions and to inmates of prisons and borstal establishments


Tables were published in the Report on Usual Residence and Workplace for England and Wales; the only tabulations for Scotland appear in the County Report series and General Volume .


The questions concerning usual residence which appeared on the census schedules were similar to those asked in 1951. The notes on the private household schedules explaining the questions however did have certain important differences.

Changes in definition

School-children and students who lived away from home during term time but were enumerated at home at the census, who had been told in 1951 to give their school, college or lodgings as their usual residence, were now instructed to give their home address as their usual residence. In a similar way, members of Her Majesty's armed forces who were enumerated at home, who had been told in 1951 to give their barracks, quarters, station or shore establishment as their usual residence, were now instructed to give their married quarters or other home address as their usual residence. These changes were made because the experience in 1951 had shown, possibly for reasons of family affinity, heads of households containing such people returned them as being resident there if they were present on census night. It was thought that the practice adopted in 1961 would be more likely to be followed and these changes were generally believed to have improved the quality of response for these groups.

The usual residence rules for inmates in institutions also differed between 1951 and 1961. In 1951, 'inmates and patients' who were expected to be discharged within six months from census day were not to be regarded as resident at the institution. The person responsible for completing the schedules for certain institutions, particularly mental hospitals, had found this rule rather difficult to operate. In 1961, the problem was approached in a different way. Institutions were divided into four groups, and the allocation of the answer to the usual residence question depended upon the group to which the individual institution belonged.

Changes in procedure for coding

The amount of coding for 1961 was much less than for 1951 as the workplace question only appeared on the ten per cent sample schedules and the system of exchange of postcards was discontinued. Instead, these duties were shared between Census Officers and the Census Office. Each officer received a list of area codes for local authority areas in and surrounding his district and also area codes for places where large numbers of people were employed. Census Officers were further instructed to code with the letter 'X' addresses which were within the local authority area of enumeration. Any addresses of which they had any doubts were to be left uncoded.

'De facto' population

There was, on the ten per cent sample schedules, an additional set of questions (Part III) for people normally resident in the household but absent from home on census night. These answers were only used for household composition analysis which is on a 'de jure' basis.

As in the past all other tabulations stated to be 'by usual residence', whether as the separate subject here or as the universe in other tabulations (subject to certain exclusions depending on the particular universe being counted), are based on the 'de facto' population adjusted for people enumerated away from home but with a usual residence elsewhere in England and Wales (or Scotland in Scottish tabulations).

Estimated resident population

For the first time however an attempt was made, in usual residence tabulations, to improve the response on the census schedule by more than the adjustment for people absent from their usual residence.

Figures are given, in Table 1 , for each local authority area, of the estimated resident population (census definition) .

This is made up of three elements:

  1. The civilian resident population based on the statement of usual residence on the census schedules. This includes in the area where they were enumerated people who said that they were usually resident outside England and Wales.
  2. Members of the Armed Forces enumerated at stations or resident in married quarters within the area.
  3. Members of the Armed Forces who had homes within the area and who were enumerated elsewhere in England and Wales outside stations. Where it seemed that substantial numbers of non-civilians absent on leave from stations in an area had (wrongly) given the station instead of their home address 'as their usual residence, the figures were adjusted to correct that error.

This definition differs, however, from the definition of resident population that is used for the Registrar General's annual population estimates and figures are also given for the estimated resident population (estimate definition) which were obtained by making further adjustments to the estimated resident population (census definition) as follows:

  1. deducting members of the Armed Forces enumerated at stations in the area and those resident at married quarters in the area;
  2. adding the estimated numbers of members of the Armed Forces who were stationed in the area at census date (those in married quarters being treated as stationed there);
  3. adding students and children usually resident at colleges, schools and university lodgings in the area;
  4. taking out of their home areas the students and children included in (iii) and also those members of the Armed Forces who were enumerated in the area and who were away from their stations or married quarters. The numbers to be removed from each area are not known and have to be estimated arbitrarily. The assumption of an even proportional spread is unlikely to produce serious errors.

Where it seemed that the school instead of the home address had been given in error as the usual residence of substantial numbers of children enumerated at school the figures were usually corrected by reducing (iii) and not by adjusting the estimated resident population (census definition).

This estimated resident population (estimate definition) is comparable with the annual mid-year estimates.

Completeness of coverage

The results of the post-enumeration survey show a small net under-enumeration but coupled with the size of the probable sampling error in the survey itself the only valid conclusion is that there is no evidence of serious errors in coverage.

There is however particular evidence of under-enumeration of West Indians. For details see 4.5.3.

Quality of response

The only important evidence from the post-enumeration survey is for people with more than one residence. Table 15 in the General Report shows the 121 people in the sample by where they normally spend the week and weekend and these results make it clear that further instruction is needed to clarify the position.

Sampling error and bias

As the tabulations for usual residence are based on the 100 per cent count they are not affected by sampling error or bias. But it should be remembered that for any subject, covered only in the ten per cent sample, where the tabulations are based on area of usual residence the chances of error must be taken into consideration.


A separate volume containing only tables on usual residence was published for England and Wales and a similar volume in Scotland.

In the Migration tables for both countries statistics are given for area of present usual residence by area of former usual residence and vice versa; and for duration of residence.


The only change in the layout of the sample census schedule was to include for all people a question on whether present or absent on census night immediately preceding the question on usual residence. But, as in 1961, only those people showing themselves to be present on census night were used in tabulations by usual residence.

Change in population included

But there was a major change in the population included; persons enumerated in Scotland claiming a usual residence in England and Wales were included in the resident population of England and Wales. Similarily those residents of Scotland enumerated in England and Wales were counted as residents of Scotland.

Estimated resident population - England and Wales

Other than the change to Great Britain as the universe, the resident population (census definition) was made up of the same elements as in 1961. No figures however were given, for areas in England and Wales, of the resident population (estimate definition). The 1966 Census being a sample census the figures are subject to various kinds of errors, the total extent of which is not calculable for any one local authority area. Sampling errors alone are, in all but the very large local authority areas, of the same order of magnitude as the adjustments necessary to produce estimate definition figures, and any deficiency in the census figures for a particular local authority area arising from defects in the sampling frame cannot be measured from the information available from the census itself; and, in the light of these considerations, it was decided not to produce estimate definition figures.

Differences in Scotland

For Scotland, in Table 1 of the usual residence tables (which were combined with birthplace to form one volume Usual Residence and Birthplace Tables '), the resident population (estimates definition) was retained.


As for 1961 a separate volume relating to usual residence was published for England and Wales and a similar volume for Scotland ; and again, in the Migration tables statistics are given for area of present usual residence by area of former usual residence and vice versa.

Office of Population Censuses and Surveys/General Register Office, Guide to Census Reports: Great Britain 1801-1966 (London: HMSO, 1977) Crown Copyright. The Office of National Statistics has granted the Great Britain Historical GIS Project permission to computerise this publication and include it in this web site. All other rights reserved.

Next Selection Previous Selection