History of the Census
Next Selection Previous Selection
The first proposal to Parliament that a population census should be taken is to the credit of Mr Thomas Potter, member for St Germans in Cornwall; on the 30 March 1753 he brought in a Bill 'for taking and registering an annual account of the total number of people, and of the total number of marriages, births, and deaths, and also of the total number of the poor receiving alms from every Parish and Extra-parochial Place in Great Britain'. The Bill was violently opposed on the grounds that it would be costly and impracticable, and that it might even be used as a basis for new taxation and for conscription. Indeed, Mr.Thornton, member for the City of York, did not believe 'that there was any set of men, or, indeed, any individual of the human species so presumptuous and so abandoned as to make the proposal we have just heard. .... I hold this project to be totally subversive of the last remains of English liberty .... The new Bill will direct the imposition of new taxes, and, indeed, the addition of a very few words will make it the most effectual engine of rapacity and oppression that was ever used against an injured people..... Moreover, an annual register of our people will acquaint our enemies abroad with our weakness'. With Government support, however, the Bill passed through all stages in the Commons by large majorities. After the second reading in the Lords on 22 May 1753 it was referred to a Committee of the whole House to be convened 'on this Day Month'. The session ended before that day and the Bill accordingly lapsed. During the latter half of the eighteenth century fear that measurements of the population might reveal weaknesses, particularly an inability to mobilise adequate military forces, gave place to fear lest the population was increasing more rapidly than the means of subsistence. The work of Malthus on population and subsistence, of which the first edition was published anonymously in 1798, was very relevant to the problems of a period when great dearth prevailed in the country and Parliament was largely occupied in discussing 'the present high price of provisions'. In these changed circumstances a second Bill 'for taking an Account of the Population of Great Britain, and of the Increase or Diminution thereof, introduced on 20 November 1800 by Mr.Abbot, member for Helston, Cornwall, passed through all its stages without opposition and received Royal Assent on December 31st of the same year. The precedent set by the Population Act of 1800 was followed regularly at intervals of ten years up to 1910, successive decennial censuses being governed by separate enactments. In the Census Act of 1920 Parliament made provision for future enumerations as well as for that due to be taken in 1921. It enacted that: 'it shall be lawful for His Majesty by Order in Council from time to time to direct that a Census shall be taken for Great Britain or for any part of Great Britain, and any Order under this section may prescribe- Provided that- [This section consists entirely of a single table listing this information.] The first census of Great Britain had two objectives. The first was to ascertain the number of persons, families and houses and to obtain a broad indication of the occupations in which the people were engaged; the second was to get information which, in the absence of data from a previous enumeration, would enable some view to be formed on the question whether the population was increasing or decreasing. The Schedule attached to the Population Act of 1800 contained the following six questions: In England and Wales responsibility for answering the first five questions was divided. The first three were addressed to those responsible for making the enumeration by house to house enquiry on the 10 March 1801, or as soon as possible after that date, a duty placed upon Overseers of the Poor (an office established under the Poor Relief Act, 1601) or 'other Substantial Householders'. The fourth and fifth questions were addressed to the Clergy. No specific reference was made in the Act to the Clergy in Scotland, where all the questions were to be answered by the Schoolmaster 'or other fit person or persons' appointed in each parish. The keeping of parish registers had been obligatory in the Church of England since 1538 when Thomas Cromwell had issued certain injunctions to the Clergy in his capacity as Henry the Eighth's Viceregent for Ecclesiastical Jurisdiction; the injunction to keep registers in proper order was repeated in the reigns of Edward VI and Queen Elizabeth. The Canons of the Church of England, which came into force in 1603, specified the manner in which entries were to be made. Inconvenience which had arisen from diverse practice in the registration of marriages was in great part overcome by the Marriage Act of 1753 (entitled An Act for the better preventing of clandestine marriages
) which required every marriage to be registered and the entry to be attested and signed by the officiating Minister, the persons married and two or more witnesses. It was for this reason that the year 1754 was chosen as the earliest for which the number of marriages was required to be given by the Census Schedule. Neither the injunctions of Henry VIII and his successors nor the subsequent Canon Law extended to the Church of Scotland. The marriage Act of 1753 did not apply to Scotland. In view of the historical importance of the first Population Act, which prescribed in detail the manner in which its aim should be accomplished, it is of interest to refer to the procedure which it laid down for the taking of the first census - particularly as the method remained substantially the same in the accounts taken of the population in 1811, 1821 and 1831. The King's Printer was instructed to send copies of the Act and Schedule to Clerks of the Peace and Town Clerks, who were required to distribute them to the Justices of the Peace within their respective limits. In England and Wales they were also required to deliver, at the Epiphany Quarter Sessions, a sufficient number of copies of the Schedule to High Constables or 'other proper Officers' as would enable the latter to ensure that one copy was received by the Overseer of the Poor or other Substantial Householder and by the Rector, Vicar, Curate or other officiating Minister in each parish, township or place. In Scotland the Schedules were received by the School master or other fit person appointed in each parish by the Sheriff Deputes or Stewart Deputes at a meeting of Justices held for that purpose. All census returns had to be made in a prescribed form. The forms given in the Schedule to the Act are reproduced on pages 42-45. Returns made by the enumerators had to be attested or affirmed before the Justices of the Peace on a day which they were authorised to fix between dates specified in the Act. In England and Wales the returns were then handed to the High Constables or 'other proper Officers' to be endorsed and submitted, together with a complete list of the names of the enumerators, on or before 8 May to the Clerks of the Peace or Town Clerks by whom they were to be sent to the Office of the Principal Secretary of State for the Home Department not later than 15 May. There they were to be 'digested and reduced to Order by such Officer as such Secretary of State shall appoint for the Purpose'. Returns compiled from the parish registers had to be forwarded by the Clergy to the Bishop of the Diocese, who was required to send them to his Arch¿bishop. Their final destination was the Privy Council, which they were intended to reach by 15 May. In Scotland the Justices of the Peace were required to forward the returns to the Home Office by 10 November. The rigid timetable set by the Act envisaged that abstracts of the returns would be laid before Parliament within six weeks of the date on which they were due to reach the Home Office. In view of the magnitude of the task it is not surprising that the abstracts were not completed until much later. The preparation of the abstracts appears to have been assigned to John Rickman[i]
, who signed the reports issued after the three censuses which followed that of 1801. The abstracts in which the results of the first census were reported[ii]
comprised an The County totals under the same heads are given for England, Wales and Scotland separately in summary tables. A general summary for Great Britain gives figures for those excluded from the enumeration, ie the Army (including the Militia), the Navy (including the Marines), seamen in registered shipping, and convicts on board the hulks. A brief statement printed below the general summary notes that (a) the total population of Great Britain must have exceeded the number of persons given in the table because no returns were received from some parishes; (b) the Islands of Guernsey, Jersey, Alderney and Sark, the Scilly Isles and the Isle of Man were not included in the census. A separate table for the Metropolis, which comprised the parishes in the cities of London and Westminster 'within and without the Bills of Mortality', appeared as an Appendix to Part I.[iii]
The Although the organization of the census was not materially different from that adopted in 1801, changes were made in th questions included on the schedules. In 1811 two important changes were introduced. The question concerning uninhabited houses was divided in order to distinguish the number of houses being built from those uninhabited for any other reason, eg dilapidation; the distinction was intended to give an indication of the degree of prosperity, or otherwise, of the district. The other alteration resulted from a decision to obtain information about families
, instead of persons
, engaged in occupations. At the census of 1831 a further question sought more detailed particulars of the occupations of the male population aged twenty years and above. The first attempt to obtain an analysis of the population by age was made in 1821 when enumerators were given discretion to record statements of age in quinquennial groups. Because of the further details required in 1831, enumerators were given more instructions than on previous occasions. Special sheets were issued for their assistance 'such as may be used in proceeding from House to House on 30th Day of May next, and on the Days immediately subsequent thereto, if one day shall not be sufficient, and by means of this Formula the account will be readily taken (in hard black-lead pencil or ink) by marks across the several lines, thus: such account to be summed together afterwards for insertion in the Schedule, by dividing it into Tens for Counting, thus: In proceeding from House to House, be careful to carry the printed Formula papers in a Pasteboard or other convenient Cover; and if ink is used by the enquirer, let him also use Blotting Paper.' The Enumeration Abstracts
presented to Parliament after each of these three censuses were compiled on the principles adopted in 1801 with the addition of preliminary observations in which Rickman explained methods used in the conduct of the census, commented on the results and gave an account of the origins of ancient divisions of England and Wales. Particulars of the numbers in Local Militia which had been embodied for exercise and training on 27 May 1811 are given at the end of each county summary in the A further important feature incorporated in the 1831 Changes made in the scope of the enquiry in 1831 led to delay in presenting the final reports which were not ordered to be printed until 2 April 1833. A summary report, entitled The 'Act for the better regulating and preserving Parish and other Registers of Births, Baptisms, Marriages and Burials in England', commonly known as the Parish Register Act, 1812 which provided for the keeping of Register Books of public and
private baptisms, marriages and burials solemnized according to the rites of the Established Church of England by the officiating Minister of every parish, did not have any great effect on the returns in 1821. The distinct mention, however, of the registry of baptisms 'whether Private or Public' added slightly to their numbers. This precluded any inference which might otherwise have been drawn from a comparison of the respective numbers of these events which took place before and after the end of the year 1812. The Act provided that special register books, one each for baptisms, burials and marriages, should be furnished by the King's Printer. An entry was to be made, by the officiating Minister, of every burial or baptism, whether private or public, within seven days of the ceremony. Special headings were provided in the books and each entry was to be numbered consecutively and a line drawn underneath. A certificate of any ceremony performed elsewhere than in the precincts of the Parish Church was to be transmitted to the nearest officiating Minister holding a Parish Register by the Minister who performed the ceremony. The entry was then copied into the appropriate register book, copies of which were to be made annually, and sent to the Registrars of each Diocese; the originals to be kept in a dry well-painted box either at the home of the officiating Minister or in the church or chapel. It was specially provided that nothing in the act should repeal any part of the Marriage Act, 1753, since the registry of marriages made under its provisions had been found most useful throughout the preceding Population Acts. In spite of the inclusion of births in the title of the Act no definite reference to their registration seems to have been made in its provisions. The effects and rulings of the Parish Register Act, 1812 extended to all cathedrals, churches and chapels (even those that were not parochial) and this increased the number of returns made under the Population Act. One important part of the act required that all officiating Ministers should send to the Registrar of their Diocese a list of all registers in their parish or chapelry stating the date of their commencement and termination, the periods during which they were deficient and the places where they were deposited. The results, however, were not encouraging and, perhaps for this reason the following question was inserted in the schedule annexed to the Population Act, 1831: 'Referring to Section XIX of the Parish Register Act of 1812 (inserted in the Register Book of Baptisms), Be pleased hereunder to insert, or to affix, a List of all the Register Books of Baptisms, Burials and Marriages (whether bound or otherwise) remaining in your Parish or Chapelry containing Entries anterior to the Year 1813; stating the periods at which the several Registers respectively commence and terminate; and the periods (if any) during which they are deficient? - If you retain Copy of such List transmitted to the Registrar of the Diocese in June 1813, a Transcript thereof would be satisfactory on the present occasion.' The lists of these registers and the details of the information contained in them were published under the tables for each hundred in the Except that the Clergy in England and Wales were only required to give an account of the number of baptisms, marriages and burials recorded in the ten years preceding each census, the form in which the returns were made and presented in the Parish Register Abstracts
remained the same for On the occasion of the 1831 Census, however, their task was made heavier by the addition of three questions. The first, which required a list of registers to be returned has already been mentioned. The second requested the ages of deceased persons registered in the several years 1813 to 1830, both inclusive. The third required a return to be made, according to the best information obtainable, of the number of male and female illegitimate children born in the parish or chapelry during the year 1830. Copious instructions were issued for the satisfactory completion of these returns. References to the registers held in each parish were given with the tables in the Two points should be noted about the 1831 From the returns of baptisms, burials and marriages compiled from the parish registers in 1801 Rickman attempted to make an estimate of the population of England and Wales at each decade from 1700. The method which he used in his calculations is described in detail on pp.xxvii-xxix of the 'The best Statement which can be given of the progressive Population of England and Wales is here subjoined, on the authority of Mr Finlaison, of the National Debt Office, who is engaged in a sedulous investigation of the expectancy of human life, from infancy to old age, founded on the materials herein explained, after subjecting them to all the tests furnished by the present state of Physical and Statistical knowledge.' In 1836 Rickman sent a circular letter to the officiating Ministers of parishes in which it was known that old registers had been preserved, requesting them to furnish him with details of the number of baptisms, burials and marriages registered during the years 1570, 1600, 1630, 1670, 1700 and 1750 and during each year preceding and following them. From these figures, supplied voluntarily by the clergy, he calculated a population estimate for each county at intervals from 1570 to 1750, using as his starting figure the enumerated county population of 1801. The detailed table of his results is printed on pp 36 and 37 of the preface to the [i]
John Rickman appeared in the Imperial Calendar
for 1814 as Secretary to the Speaker; in 1815 as Second Clerk Assistant in the House of Commons; and in 1822 as Clerk Assistant. A footnote on page 4 of the General Report on the 1901 Census contains a description of John Rickman given in a letter written by his friend Charles Lamb. [ii]
Their full title was Abstracts of Returns and Answers made pursuant to an Act, passed in the Forty-first Year of His Majesty King George III entitled 'An Act for taking an Account of the Population of Great Britain, and the Increase or Diminution thereof . [iii]
When attempting to correct the lists of parishes in this table, Rickman appears to have confused St. Botolph Billingsgate with St. Botolph Bishopsgate and, in consequence, the first two items of the errata given below the Summary of Enumeration at the beginning of the Abstracts
are incorrect. [iv]
0n the question of the validity of these calculations, which is not dis¿cussed here, it should be noted that Rickman's methods were scarcely in harmony with those of, eg Joshua Milne who was his contemporary. [v]
Reference to these estimates was made in a paper on the population of England in the 18th Century read by E.C.K.Conner in 1915 and published in the Journal of the Royal Statistical Society, Vol LXXVI, pp 261 et seq. By the Population Act, 1840 the Registrar General, and such other persons as should be associated with him by Her Majesty, were constituted Commissioners for taking account of the population of Great Britain. 'The Census of 1841, so far as it relates to England and Wales, having been taken on a plan never before attempted and by means of machinery which was not in existence at the period of any previous Enumeration of the People, it may be proper at the commencement of this Paper to notice that the Act of 3 and 4 Victoria Cap.99 by which it was authorised was framed upon a scheme propounded to the Government in June, 1840 by the late Mr.Lister, the then Registrar General of Births, Deaths and Marriages, the details of which will appear in the following Letters addressed by him to Mr.Drinkwater Bethune, the Counsel employed by Government to prepare the Bill'. It is evident from the copies of the letters reproduced in the manuscript document preserved in the General Register Office, to which the passage quoted above forms the preface, that the first Registrar General planned the details of the 1841 Census with the same care and thoroughness that characterized the organization of the civil registration of births, marriages and deaths which he had accomplished in 1837. The method of conducting the enumeration on that occasion has stood the test of time and has not been substantially altered since. In order to avoid omissions on the one hand or, on the other, the counting of heads twice, the Registrar General's first aim was to ensure that the census was taken everywhere at the same, and in the shortest possible, time; on one day or, at the most, two. The whole of England and Wales was already divided into 2,193 Registration Districts, the limits of which were known and published. For the purpose of the census each of these was sub-divided into an appropriate number of Enumeration Districts. Plans of Division were prepared by each registrar and, after examination by the superintendent registrar, sent to the Census Commissioners. The results of a trial census held in three different types of locality towards the end of 1840 assisted in determining the criteria laid down in instructions to registrars. Enumeration districts were necessarily limited in size by the varying circumstances of ease of communication or density of population, so that they should contain not more than 200 and not less than 25 inhabited houses. In thinly populated areas where houses were scattered the district included no larger an extent of country than could be covered by an active enumerator in one day. As a consequence of the small size of some of these districts 35,000 enumerators were required in order to cover the whole of England and Wales. Arrangements were made for all public institutions, barracks, gaols and workhouses to be enumerated by the officers in charge of them. The original act, which was later amended, made provision for the census to be taken on Thursday 1 July. The enumerators were to carry out much the same function as on former occasions, but they were under the direct supervision of the registrars. At each of the first four enumerations a Monday had been fixed for the census day on the probable assumption that more people would be at home on a Sunday night. John Rickman, however, in his Preface to the Under a further provision of the later act schedules were, for the first time delivered to every householder throughout England and Wales and Scotland, a few days before the appointed day of census. Each householder was directed to complete the form in respect of all persons sleeping in the house on Sunday 6 June, before the arrival of the enumerator on the following day, the penalty for failure to comply with these instructions being a fine of not less than 40 shillings and not more than ¿5. This innovation reduced the chance of omissions or double entries and assisted the endeavour to complete the numbering of the people in one day. The task of the enumerator after census day was to ensure as far as possible the correct and complete return of the schedule. Later he had to transfer the answers to his own schedule. The information he had to copy was not so simple as formerly, since under the new method an account was taken of each person individually. The required details were entered under the appropriate column opposite each person's name, and the enumerators did not merely enter the numbers in each household as had previously been the case. The enumerators' schedules, after being examined by the registrars and sub¿mitted to, and counter-signed by, the superintendent registrars, were then sent to the General Register Office. It was found that under this new system every enumerator was able to fulfil the aim of the new procedure by collecting his householders' schedules within the day. Civil registration was not established in Scotland until 1855 and superintendence of the census was again entrusted to the official schoolmaster or other fit person in each parish. The same preliminary measures were taken, however, in Scotland as in England and Wales, each parish being divided into enumeration districts, which could be conveniently covered in one day by the enumerator. The Sheriff Substitute was responsible for examining and signing the schedules in the same way as the superintendent registrar in England. The questions on the schedule were more extensive and rather more detailed than in 1831. In addition to the name, sex, age and occupation of every living person, householders were also asked to state which persons were foreigners, and which were born in the parish or county in which they were living. The occupation of every person, regardless of age or sex, was required to be stated other than wives or sons or daughters, living with their husbands or parents and not receiving wages. At each census from 1801 to 1831, enumerators had been instructed to take an account of the number of persons actually found within the limits of each parish, township or place, 'exclusive of men actually serving in His Majesty's Regular Forces or Militia, and exclusive of seamen, either in His Majesty's service or belonging to registered vessels.' In 1841 the instruction was different. Soldiers and sailors ashore in Great Britain were enumerated at their barracks or places of residence and entered under the parishes in which they slept or abode on the night of June 6. 5,016 persons were returned as travelling on railways and could be attributed to no particular locality. Although the unit of enumeration was based on registration districts at this census, information given in the reports was presented as previously for the traditional divisions. A greater effort was made to distinguish the population of more of the distinct places forming parts of parishes, such as townships, tythings, hamlets or villages. The The data collected were published in three volumes, the The The 1851 Census opened a second half-century of census history. The general machinery and organization remained unchanged, but the scope of the enquiry was greatly extended. The householder's schedule required in respect of each person a statement of relationship to the head of the household, of condition as to marriage (ie whether unmarried, married, widow or widower) and whether the person was blind or deaf and dumb. Exact age was also to be stated; in 1821 and 1841 it had been asked to the nearest quinquennial age-group. (Details of ages by single years were not published, however, until the 1911 Census). The coverage was also extended. Provision was made to enumerate all persons on board vessels lying in harbours and navigable rivers; they were returned on a schedule completed by the Master of each vessel and handed by him to Officers of the Customs who acted as enumerators. In addition to the census of those at home, the Admiralty and the Registrar of Merchant Seaman also carried out an enumeration of those at sea in ships of the Royal Navy and the Merchant Service. Similar arrangements were made in respect of the Army abroad, of Europeans in the service of the East India Company and of British Subjects residing in various foreign states. The fruits of experience gained in 1841 were reflected in the instructions printed on the householder's schedules and in the more detailed instructions issued to Registrars and enumerators. Particulars of the schedules, forms and instructions issued to Registrars and enumerators were published in a Command Paper presented to Parliament on the 14 March 1851 (ie a fortnight before Census day); details of the more important among them are also given on pages cxlii-clvi of the Report in the first volume of A fundamental change was made in the form in which the census tables were presented. The Registrar General had already, for statistical convenience, grouped the registration counties of England and Wales into eleven Registration Divisions. The area, population, topographical position, historical connections and occupations of the people had been taken into account when the divisions were drawn up for use in the tables of vital statistics published in the Registrar General's Annual Report
. In the census tables, which related to England and Wales, statistics for each parish or place were arranged under their respective registration districts and sub-districts grouped by registration divisions. Scotland was divided on an analogous basis into two groups of counties and the Islands in the British Seas formed a fourteenth group. Particulars of the counties included in each of the registration divisions of England and Wales and the northern and southern portions of Scotland are given on page 263. The boundaries of these division are shown on maps published in the first volume of The tables comprise two series. The Reports might be considered the most interesting in the census series from 1801 to 1931. Signed by George Graham, the Registrar General, and by William Farr and Horace Mann, his assistants in the matter of the census, they throw considerable light on contemporary thought on social problems as well as containing much of historical flavour. The several charts and maps with which they are illustrated include maps depicting the density of population in England, Wales and Scotland An Appendix to the Report in the first volume of Two important enquiries associated with the 1851 Census were those into religion and education. Plans were made on the assumption that these enquiries could be conducted under powers given in the Census Act of 1850. Objections to the penalties which could be imposed on persons withholding information on these subjects, raised in the House of Lords, were later upheld by the Law Officers of the Crown. The Registrar General decided to go ahead with the surveys on a voluntary basis, care being taken to inform every person concerned that answers to the questions on the special schedules was not compelled by law. There were few refusals. An account of these surveys and of the information published in the Reports upon them - which were not issued as part of the official census series - is given on pages 204 and 217. In addition to the preliminary statement presented under the title In addition to summary tables for England and Wales and the Islands in the British Seas (Nos 1-146), the Appendix had a table showing the areas of parishes and townships in four northern counties, ie Durham, Northumberland, Cumberland and Westmorland (No 333) and a return of the number of inhabited houses and the population at 1861 in the several wards formed for the election of vestrymen in metropolitan parishes divided under the Metropolis Local Management Act of 1855 (No 334). The Appendix concluded with two very interesting papers by William Farr. In the first he discussed the nature of products and producers, the etymology of the names by which various occupations are known, the value to the community of those pursuits followed by the professional classes, and briefly, the occupations of those comprised in each of the six classes of the new occupational classification which he had prepared for use in connection with the 1851 Census. The second paper consists of a letter to the Registrar General on a proposed enquiry into the occupations of the people with an outline of the heads of enquiry, a sample reply on coal mining in Staffordshire, a short memorandum on the Cornish mines and a draft report on the medical profession. From the heads of enquiry and the sample reply obtained from the manager of a South Staffordshire colliery, it is evident that this proposal would have provided information of a kind since obtained through the Census of Production. The civil registration of births, deaths and marriages began in Scotland in 1855 under the control of a Registrar General for Scotland to whom was given, under a separate Act passed in 1860, responsibility for taking the census of 1861 in that country. Registrars were required to divide their districts into convenient units for the purpose of enumeration. The registrar's Plan of Division
, was forwarded to the Sheriff of the County or the Chief Magistrate of the Burgh, as appropriate, for approval together with a list of the enumerators proposed for the various divisions. Public institutions with upwards of fifty inmates were treated as separate enumeration divisions, the officer in charge having been appointed as enumerator. In general the organization, which is described in the first volume of the The only change made in the information sought at the 1861 Census was that in Scotland particulars were obtained of the number of rooms with one or more windows and of the number of children between the ages of 5 and 15 who were attending school or being educated at home. Apart from these two questions, the census enquiry was the same throughout Great Britain. Following the administrative changes mentioned above, the 1861 Census marked the first occasion on which a separate Report
on the Census of Scotland was presented to Parliament. It comprised two volumes; the The questions asked at the 1871 Census only differed substantially from those of the preceding English and Scottish censuses by a request that the word 'Unemployed' should be added to the occupational description of persons ordinarily engaged in some industry but out of employment on 3 April. This was repeated in 1881, but no reference to unemployment was made in any sub¿sequent census until 1931. In 1891 three additional columns were added to both English and Scottish schedules to distinguish employers, employed and those working on their own account. The schedule used in England and Wales on this occasion introduced a panel with the following instruction: 'If you occupy less than five rooms write in this space the number of rooms occupied by you'. A count of the gaelic-speaking population was first made in Scotland in 1881; the enquiry into the languages spoken in Wales was introduced in 1891. While the pattern of the Scottish Reports remained unaltered, the English tables of area, population and housing were divided between two volumes. The first presented data for ancient counties, arranged alphabetically with their respective parliamentary divisions and boroughs; hundreds, wapentakes, etc; boroughs, principal towns, sub-divisions of lieutenancy; parishes and places in 1861, with the addition of ecclesiastical parishes and petty sessional divisions in 1871 and 1881, and of administrative counties in 1891. The second gave, in registration division order, comparable figures for registration counties and registration districts and parishes within them; in 1861 data were also given for ecclesiastical parishes. Details of ages, civil condition, occupations, etc were published in a third volume. As in 1861, the General Report
for England and Wales was published separately on each occasion. Comparisons with French statistics were a feature of the Of the Reports on the Censuses of Scotland
published during this period, the With the passing of the Census (Great Britain) Act, 1900, Parliament returned to the practice, changed in 1860, of making provision in a single enactment for the taking of the census throughout Great Britain. The general scope of the enquiry was practically the same as in 1891, but an addition was made to questions on occupation in order to ascertain the number of people who carried out their trade or industry at home. Although the form of the Scottish Report
was not altered, the tabular matter for England and Wales was published on a different principle from that which had been used since 1851. The whole of the information pertaining to each county was incorporated in a separate The period between 1891 and 1901 was characterized by very considerable revision of the boundaries of local government areas. The extent to which this increased the task of the Census Commissioners can be judged from the number of some of the changes. Of the 14,900 civil parishes for which populations were given in the 1901 census tables, 883 were created between 1891 and 1900 and 1,308 had their boundaries altered. Urban districts had been created during the decennium to the number of 164; 53 were abolished and merged into other districts and 281 underwent alteration. As the result of changes made under provisions of the Local Government Acts of 1888 and 1894, less than a third of the rural districts remained the same as in 1891. In addition, the tables reflected the division in 1899 of the Administrative County of London into 28 metropolitan boroughs. The most important innovation made in 1911 was the use of the census to obtain more detailed information bearing on the fertility of marriage. For this purpose the householder's schedules required particulars to be given of (a) the duration of existing marriages, and (b) the number of living children born to each marriage and the number alive at the date of the census. The field of enquiry was also extended under other heads. Infor¿mation was required (a) about the industries or services with which workers were connected, ie as distinct from the occupations in which they were personally engaged; (b) whether persons born outside England, Wales or Scotland were residents or visitors; and (c) for England and Wales, the number of rooms in all dwellings and not only, as at the two previous censuses, in dwellings of less than five rooms. The 1911 Census was the first in which machines were used for purposes of tabulation in Great Britain. From 1841 to 1901 it had been part of the duties of
each enumerator to copy the replies to the questions on the house¿holder's schedules into his Enumeration Book from which the census tables were then prepared. The introduction of machine tabulation made it necessary to code most of the particulars and this was done direct from the schedules themselves; the process of copying was omitted and the chances of error therefore lessened. On completion of the coding, the schedules passed into the hands of machine operators who recorded the coded information on special cards by punching holes in appropriate numbered positions. Particulars relating to any one individual were recorded on 'personal cards' and material needed for the fertility analysis was recorded on a second set. A third set was used to assemble, from summaries made by the numerators, details of population and buildings in each enumeration district. This process was followed by machine tabulation carried out in two stages by means of two other machines which (i) sorted all the cards with holes in certain identical positions and (ii) counted the cards thus sorted. A full description of these processes was given in Appendix B to the 1911 Another change in procedure, made in England and Wales, was the careful preparation of a series of maps designed to ensure accuracy in the assign¿ment of population and houses to various areas for which tables were to be published and to assist in revision of the Plans of Division, ie the sub¿division of registration districts into enumeration districts. Preparation of a preliminary set of maps, begun two years in advance of the census, was carried out by reference to local maps and records which were borrowed for the purpose. With the co-operation of the Ecclesiastical Commission and of Diocesan Registrars the boundaries of all the ecclesiastical parishes of England and Wales were charted. The preliminary set was then sent to the Ordnance Survey Department where a second set of maps was prepared for each registration sub-district showing in distinctive colours the boundaries of civil parishes; urban districts; municipal boroughs; wards of urban districts; municipal boroughs and parliamentary divisions; and ecclesiastical parishes. A significant consequence of the introduction of machinery was that it enabled the results of the census to be presented in greater detail than could ever have been attempted before. The final reports for England and Wales extended to sixteen separate volumes, of which The Scottish Report
, although more detailed than before, did not follow the plan adopted in England and Wales. Instead, the publication of the The 1921 Census was the first to be taken in accordance with the provisions of the Census Act, 1920, which, unlike previous Acts governing the census, was a permanent enactment applicable to all future censuses in Great Britain. It had been intended that the census should be taken on the 24 April, but owing to conditions occasioned by a dispute in the coal mining industry and the expectation of a strike of railwaymen and transport workers it was decided to postpone the enumeration until the 19 June. The date of the census, the persons by whom and with respect to whom the returns were to be made, and the nature of the particulars to be furnished in those returns were prescribed by an Order in Council made on 21 December 1920, which was later amended in respect of the date of the census. Regulations under the Order were made at the same time. In the matter of procedure at the taking of the census two changes were made. For the first time provision was made for separate confidential returns to be made by persons who would otherwise have suffered hardships by disclosing particulars to some other person charged with the duty of making the return. Arrangements had, of course, to be made to link such separate returns later with the household schedule in which they would normally have been included in order that statistics based on the household unit might not be affected. The other change was that registrars were instructed direct from head¿quarters and not, as on previous occasions, through Superintendent Registrars. The regulations provided, however, for the appointment in each area of a Census Advisory Officer; with few exceptions, these appointments were accepted by Superintendent Registrars who acted in an honorary capacity. When preparations for this census were begun towards the end of 1919 particular emphasis was laid on the desirability of securing co-ordination in the census enquiries to be undertaken in all parts of the United Kingdom. To this end a Census Joint Committee, consisting of the three Registrars General responsible respectively in England and Wales, Scotland and Ireland, was constituted. Matters proposed for inclusion in the schedule were grouped according to subject matter and submitted to expert sub-committees for examination in the first instance. The schedule annexed to the Order in Council made on 12 February 1921, replaced two subjects included in 1911 by three new questions. The enquiry into infirmities, which it had long been felt did not furnish reliable information, was dropped. The special enquiry into the fertility of marriage made in 1911 was omitted, not because it was considered unimportant, but because it was decided that the balance of advantage lay in providing for the inclusion of two new enquiries. It had been emphasised in the 1911 It is believed that the question as to place of work, introduced in England and Wales in 1921, had not previously been asked at any other census in the world. The importance of this enquiry arose from the need to obtain some measurements to assist in resolving problems of traffic, transport and housing which inevitably flow from the tendency, in a highly organized and industrial community, for residential areas to become separated from manufacturing and commercial centres. The enquiry into place of work was not made in Scotland; instead, it was required to be stated whether those enumerated were entitled to benefit under the National Health Insurance Acts. The second new field of enquiry related to the ages and numbers of children under 16 and into the number of orphans among the enumerated population. The aim was to ascertain the extent of the burden of dependency in different sections of the community. The third new requirement in England and Wales was a statement whether those attending school or any educational insti¿tution for the purpose of instruction were doing so full-time or part-time. A minor adjustment made to the schedule was a requirement that age should be stated in years and months in order to lessen mis-statements of age. The 1921 Census saw a return in England and Wales to the publication of The Report on the Census of Scotland
followed the pattern of 1911; orphanhood was added to the subjects covered in The first part of the The enquiries into workplace, orphanhood, dependency and education were not repeated at this census. The schedule, drawn up after due consideration of various proposals submitted, contained one new feature, a requirement that place of usual residence should be stated. In Scotland the question regarding entitlement to medical benefit under the National Health Insurance Acts was repeated and length of residence was asked in respect of persons not born in Scotland. Reports published after this census followed the pattern of those issued for 1921, except that a separate volume was given to This enumeration was carried out in the United Kingdom and the Isle of Man on 29 September 1939 for the purpose of compiling a register of the population as a war-time general security measure. Certain statistics from the enumeration were published in 1944. The procedure by which the initial register was compiled in Great Britain and the Isle of Man differed in certain material respects from that used in Northern Ireland and both are noted below. Preparations for the National Register were begun at the end of 1938 and were completed by April 1939 except for the recruitment of enumerators, which continued in some areas until shortly before the outbreak of war. If the country had stayed at peace the arrangements would have held good for the projected 1941 Census. The enumeration districts were certainly smaller than usual, but this was necessary as Identity Cards had to be issued in addition to the normal delivery and collection of schedules, and completion of the whole operation was required as soon as possible. In England and Wales, planning was carried out as before by registrars of births and deaths, but supervision of the enumeration was entrusted to the Clerks of Borough, Urban and Rural District Councils, having been appointed for the purpose as National Registration Officers. In Scotland, the registrars handled the whole operation and in the Isle of Man it fell to Town Clerks, Clerks to the Commissioner and to Captains of Parishes. Registration was compulsory for all persons present on National Registration day except those serving in and not on leave from HM Forces. The information to be recorded was the minimum needed for the purposes of National Registration and comprised name and address, sex, date of birth, marital condition and occupation. In the enumeration processes, census practice was generally followed in that the basic unit was the household schedule, with other types of schedule being issued for completion by managers of hotels or resident officers in charge of institutions. Records of more than 45 million persons in 65,000 enumeration districts were produced. As the original schedules were reserved for the use of local food authorities as a basis for the preparation of ration books and to constitute the local Food Registers, enumerators were instructed to make transcripts of the schedules before parting with them. Special Transcript Books had been prepared and issued for this purpose and after completion, were checked by local registration officers before being sent to the central offices. Also at this time each enumerator was instructed to complete a form of statistical analysis of the population content of his area. This was based on the data entered in the Transcript Books. Local registration officers were again called on to check (only a sample on this occasion) the forms which were further scrutinised and tested at the central offices before the final aggregation was made. All data contained in the published volume were prepared by this method. The National Registration population figures differ in one important respect from those of a normal census, which aims to minimise temporary movements of population (eg industrial or other holidays) as far as possible; and for that reason is conventionally taken in the spring. However, in the case of the National Registration enumeration, very considerable movements of the population had taken place since the outbreak of war due to evacuation, etc and these did not conform to any established pattern. Nevertheless, statistics of actual local populations, however abnormal and erratic were required and this the NR enumeration certainly provided. It can be seen, though, that the data could have no continuity with past or future censuses and hence had no standing as a basis for future population estimates. There was no independent enumeration (census type) in Northern Ireland on 29 September 1939 and the initial register was compiled from the returns obtained from the Food Control authorities for the purpose of rationing. Application forms for ration books were delivered to all householders on or about 25 September 1939, through the postal services, excluding inmates of institutions and serving members of the Armed Forces. Local food offices prepared ration books on receipt of the completed forms which were subsequently sent to the Registrar General, Belfast to form the basis for the issue of Identity cards and for the preparation of the National Register. The census of 1951 for England and Wales was wider in scope than previous enumerations and this can be attributed to two main factors: (a) the interval of twenty years since the previous census had produced many legis¿lative and social changes and (b) the census of 1931 was taken at a time of severe economic depression and was restricted to the collection of basic information. As at every census since 1841, the local arrangements for the enumeration were based on the areas covered by the local registrars of births and deaths (the registration sub-districts), and in general they acted as Census Officers. This, incidentally, was the last occasion on which, as part of their duties these officers were called upon to divide their areas into suitable enumeration districts; in 1961 and 1966, the Plans of Division were prepared centrally in the General Register Office. After the enumeration, Census Officers had certain duties to perform before despatching the main bulk of schedules and other documents to the processing office then at Southport. The first items to be handed over by the enumer¿ators were the Enumeration Books from which was prepared a summary form setting out population and other brief data for each local authority area and for the sub-district as a whole. The summaries were used in the compilation of the Thirdly came the examination of the workplace and usual residence postcards which were prepared by the enumerators as in 1921 when details of the place of work were last asked. The enumerator was required to complete postcards for each address of usual residence or workplace which was outside the local authority area in which his district was located. On each card, an address had to be written on one side (and also the employer's name on workplace cards) and on the reverse side, reference numbers (to identify the schedule), the sex of the person concerned and age (usual residence cards only) were required. On receipt of the cards, the Census Officer checked them and entered a code for the area of enumeration before handing them to the local Post Office. They were treated as ordinary mail except that on reaching their office of destination, the cards were diverted to the Census Officer for the area instead of being delivered to the actual address. Incoming cards were coded for the area of usual residence (or workplace) before being sent forward for processing. The creation of a one per cent sample was designed to provide, relatively quickly, a cross-section of data of a much wider scope than the The 1951 questions can be summarised in three groups: enquiries made in the preceding census and repeated in a substantially unaltered form, those appearing again after an interval, and those which were wholly new. The first category included sex, age and marital condition, relationship to the head of the household, birthplace, nationality, number of rooms occupied by the household, occupation, industry, usual residence and, in Wales and Monmouthshire, the Welsh language. The second group comprised place of work (last asked in 1921), whether receiving full-time or part-time education at an educational establishment (1921) and an enquiry on fertility (1911). This enquiry concerned the date of marriage of women under age 50, the date of their first marriage if married more than once, the total number of children born alive to them in marriage and whether a child had been born in the twelve months preceding census day. It was considered that sufficient information about the fertility experience of women over 50 was available from the special Family Census conducted on behalf of the Royal Commission on Population in 1946.[i]
There were two items which were wholly new in 1951. Firstly a further question on education was asked: for those whose full-time education was complete a statement of the age at which it had ceased. This question was addressed only to the gainfully employed. The other new topic comprised a question about household arrangements and enquired whether each household had the exclusive use of, or shared with another household or lacked entirely, piped water supply within the house, cooking stove or range, kitchen sink, water closet and fixed bath. An interesting point on household arrangements is that whereas with other topics it is often possible to relate the material to previous censuses or to other data in the Department (eg birth and death registrations) for checking, no such facilities existed for this subject. Therefore, in an effort to assess the quality of the data, a small sample of households in unshared occupation of dwellings was chosen in six local authority areas, and local registration officers made visits to the addresses concerned in order to verify the information previously given. (See 4.2.2). The technique of a post-enumeration survey was evolved in the USA and this study, in 1951, was the first example of its kind in this country. The scheme of publication of reports for England and Wales was broadly the same as in 1931 and 1921. New volumes, however, comprised the The information published is no more than a selection of the potential total and some special tabulations were undertaken to meet the needs of other Government Departments, Universities, etc but are not generally available. In Scotland the arrangements for taking the census were in all essentials the same as in the past. The central Census Office was set up in New Register House and in each registration district the local arrangements for organising and taking the census were entrusted to the Registrar of Births, Deaths and Marriages, acting under the instructions issued by the Census Office. Although most of the data collected were identical with that for England and Wales, the tabulation and analytical processes were not so highly developed and no General Report was published. The most noteworthy features of the sixteenth census of England and Wales were the use of a computer for processing and the use, for the first time, of sampling methods in the field at the time of the enumeration. In 1957, when consideration of a machine to process 1961 Census data began, manufacturers were invited to submit proposals. All recommended the use of a computer, with a much higher speed of operation than conventional punched card machines, and subsequently it was decided that the General Register Office would share an IBM 705 machine to be installed for use by the Royal Army Pay Corps at Worthy Down, near Winchester in Hampshire. It was later found that both the machine time required and time taken to write computer programs had been greatly underestimated. The use of a computer made it practicable for the first time to have a machine output comprising the precise figures required for publication, arranged in the way they were wanted. It would thus have been possible to produce the data in the reports directly from this--'output by photographic means, but since this print would provide neither the quality to be desired nor any variety of type founts, it was decided that the computer output should be by punched cards, to be fed to a series of card-controlled typewriters which provided combinations of type fount from which the desired presentation was obtained. Although the improvement in the timetable was not as great as was expected nevertheless twice the volume of results was produced in less time than was taken with the 1951 Census. One of the earliest tasks was planning the sizes and boundaries of the smallest areas to be identified at the census - the enumeration districts; and as the figures to be produced from the census relate to local authority areas, planning ensured that each ward or civil parish comprised one or more complete enumeration district. On previous occasions the preparation of the Plan of Division was carried out by the Census Officers, usually the local registrars of births, deaths and marriages, but for 1961 it was decided to plan centrally because of the complex nature of the work and the inability, through other commitments, of many registrars to produce an acceptable plan on time. Work on the Plan of Division for the 1961 Census commenced in July 1958 and the instructions were to plan enumeration districts to contain some 250 households in urban areas and in rural areas the 1951 districts were to be retained unless there was a good reason for a change, such as considerable housing development or changes in local authority boundaries. During the planning however it was found that some registration sub-districts would contain too many enumeration districts for efficient control. Where practicable three or more sub-districts were grouped together and then redivided into 'Census Districts' containing not more than 70 enumeration districts with about 50,000 population. The Census Regulations, 1960 differed from earlier regulations in that they did not impose census duties on Registration Officers as a whole, although most of those employed as Census Officers were Registration Officers: of 1,315 Census Officers, 1,200 were from the registration service and the bulk of the remainder were from local government. In all, almost 69,000 enumerators were appointed, the three chief sources of supply being local government officers, civil servants and housewives. Finally it was decided to appoint, in some areas, Census Advisory Officers with, unlike 1951, specific duties to perform and for which they would be paid. The post was offered to certain selected Superintendent Registrars and 115 were appointed. Their duties were to (a) recruit, interview and select enumerators and (b) answer enquiries from the press, local officers and the public. With the advent of the computer it was not considered necessary to repeat the one per cent sample as in 1951 and early in the planning stage consideration was given to the production of tables on certain topics on a sample basis only, without the repetition of a full count. The chief advantage of this system is, quite simply, economy at the processing stage both in manpower (for coding and punching) and machine time. The chief drawback of sample-based figures is their lack of precision and care is needed at the planning stage to ensure that the size of the sample selected and the size of the area for which figures are to be presented are both adequate. For 1961, it was decided that a ten per cent sample would be large enough for those topics for which statistics were to be compiled mainly at national level. Thus economic activity ( The theoretical method of schedule distribution was quite simple: People living in institutions, hotels, ships etc were not enumerated on the ordinary household schedules but on other special forms (types I and S). As the size of these establishments varied too much for an institution sample to give reliable figures, it was decided that in these 'non-private households' the sample should be of individuals. This was achieved by asking the extra questions only of people appearing on a specified line on each of the special schedules. The sample lines were also designated randomly and the person completing the schedule was given strict instructions to avoid any pre-selection of people from whom the extra infor¿mation would be required. No sampling arrangements were made for the enumeration of Home Forces personnel (on schedules NMA) for whom the sample was selected at the Census Office. The method of sampling at the enumeration stage, outlined above, was an innovation and plans were made to test the validity of the sample as finally selected. These tests were of two kinds. The first was intended to reveal any bias in the returns from individual enumeration districts or local authority areas, while the second aimed at ascertaining whether significant bias existed in figures produced at national level. The procedure adopted in both cases was similar in that comparisons were made on areas or topics which had been tabulated on both the full count and the ten per cent sample basis. In enumeration districts where a possible irregularity was discovered, the sample selection was checked by an examination of the Enumeration Record Book and, by this method, a number of departures from the intended sampling scheme were found. Another check was made, this time on Census Office procedure, in a comparison for a few areas of the sample as originally selected and in the form in which it was finally processed. No significant difference was found and it seems clear that the bias was introduced at the enumeration stage. Other than this, it was not possible to obtain any objective evidence as to the basic causes of the bias. As to the subsequent action taken, the decision was made not to amend the actual numbers obtained from the sample in the published tables, but to produce certain correcting factors which could be applied as appropriate. These factors were not produced for every entry in the tables or even for all sample tables, but for certain of the more important marginal totals only (eg each of the occupation and industry orders) because the calcula¿tion of factors for all entries would have produced a further and unacceptable delay in processing. (For Sampling error and bias see 7.2 -7.3). The longer schedule contained the largest list of questions ever asked and included two new topics: In addition, a Part III was added to the form and asked for details of persons absent from the household. This information was used in the Household Composition analysis which it should be noted was on a much larger scale and used a different method of classification from that used in 1951 (See 4.2.3). A new housing question was also included, on both types of schedule, which was designed to ascertain how the household occupied its accommodation: and there were changes in the question on household arrangements. The questions on Fertility (on both schedules) were extended in 1961 to include all women who were or had been married. The questions about children under this heading were unchanged but a new enquiry asked for the date of termination of the first or only marriage for widowed or divorced women or women married more than once. Processing was started, as is customary, by the production of data for the The cards were passed through a 'mark-sense reproducing machine[ii]
which was designed to punch holes automatically for the figures marked on the cards. Essential to its proper functioning were: (a) the cards should not be creased or damaged in any way, (b) the marks should be in black lead pencil and join the brackets enclosing each figure and (c) the marks should be made firmly and in a continuous line. Some thousands of the cards received offended one or all of these essentials and time was wasted while some 15,000 cards were re-marked. In addition, about 24,000 cards were re-punched manually because either they were damaged or had been wrongly completed by the enumerators. Finally, some enumerators were extremely slow in sending their report cards and even at the expiry of a deadline two weeks after the original date for dispatch some 300 cards were still outstanding. For these areas, estimated figures were used. It was also decided to extend the limited use in 1951 of the technique of post-enumeration survey. The aim of the survey was to obtain information on both coverage and quality of the enumeration and for obvious reasons it needed to be carried out quickly. Accordingly this operation took place during the first two weeks of May 1961. In order to do the survey at a reasonable cost and also for practical considerations it was decided that the sample of households to be used for the coverage check should itself be sub-sampled for the quality check. The sample actually used for the coverage check comprised some 2,500 plots (sub-divisions of enumeration districts) each containing about 20 households and selected at the Census Office by a random chance procedure from EDs listed in geographical order. The quality check sub-sample was obtained by taking all the households in the coverage sample which were originally enumerated on a 10% schedule (E.10) plus
one household for each plot enumerated on a 90% schedule (E.90); this gave a rough average of three households per plot. The actual inter¿viewing was carried out by selected enumerators under the direction of the appropriate Census Officers. A detailed description of the methods used and an evaluation of the results of the survey are to be found on pages 45 to 54 of the The list of publications for 1961 show a number of changes compared to that for 1951. In addition to the standard publications, volumes relating to new subjects were published. Also a series of To meet a growing demand from various universities, research organisations, etc tabulations were produced in something of the detail accorded to larger areas but relating to wards and civil parishes and are available for the cost of reproduction. Also, although a volume relating to Ecclesiastical areas was not published data are available in unpublished form. In addition to the particulars required in both countries, the Census Order provided that returns made in Scotland should state the county
of birth of persons born in Scotland. For England and Wales only country
of birth was required although county
had been asked in 1951. In the Order of 1960, complementary to the Welsh language questions there was also provision for obtaining information on persons in Scotland able to speak Gaelic. It can thus be seen that there was very little difference in the basic information to be supplied and divergencies of practice at the processing stage leading to irritating discrepancies between the published versions of tables were largely eliminated, at an early stage, by the decision to also process the 1961 statistics for Scotland on the Royal Army Pay Corps IBM 705 computer at Worthy Down. With the exceptions noted above, in fact, the same particulars were to be obtained throughout Great Britain which made it possible for programmes prepared for English data to cover Scottish needs with relatively few modifications. In Scotland the division of the country into census districts was undertaken at the Census Office in Edinburgh. With few exceptions the local registrars of births, deaths and marriages were appointed to act as Census Officers, and for convenience the boundaries of the census districts were made to coincide with the boundaries of the registration districts. For the last time each Census Officer was given the task of planning the enumeration districts for his own area, roughly on the basis that enumeration districts in urban areas should each contain about 200 households; in rural and sparsely populated districts the number of households varied considerably and a high proportion of them contained fewer than 100 households. In the whole of Scotland there were some 950 census districts and 10,400 enumeration districts with corresponding numbers of Census Officers and enumerators. There were no Census Advisory Officers in Scotland. The instructions to the Census Officers and enumerators differed in some respects from those issued in England and Wales but only where Scottish conditions made this necessary. The packs of household census schedules were compiled in the same way, a random starting number being allocated to each enumeration district to determine the position of the household schedule containing the full range of questions which was issued to one household in ten. Subject to a few modifications designed to obtain the additional particulars referred to, the forms of return were similar to those used in England and Wales. The There was no post-enumeration survey in Scotland. At the Census Office in Edinburgh the information on the census schedules was checked and coded and the data were put on punch-cards which were sent to the computer centre at Worthy Down. The process thereafter was similar to that for the England and Wales punch-cards except that any queries had to be sent to Edinburgh for clearance there. The Reports on the Scottish results were similar in scope to those published in England and Wales with a few differences, of which the following are the more important: The seventeenth census of the population of England and Wales was taken as at midnight on 24 April 1966 and on the same date in Scotland. This was the first ever quinquennial census in Great Britain and while sampling methods had been used before (in 1951 and 1961), the 1966 Census was the first occasion on which information was required only from a section of the population. The Census Act 1920 makes provision for a quinquennial census and it was established that a sample enquiry would also be within its powers. Broadly a ten per cent sample of dwellings throughout the country was constructed and the remainder had no part in the enumeration. A similar situation applied to Scotland but six special study areas had also been designated which were enumerated and processed 100 per cent. The main sampling frame for England and Wales was the 1961 census record; from this, a computer listing was obtained of: Private dwellings which were part of, or associated with, a non-private establishment were removed from the sample drawn, while caravans were also excluded. These lists were then brought up to date to allow for new building since the 1961 Census and to include residential caravans. The list of 'large' non-private establishments for each local authority area was sent to the appropriate authority who added the names of new establishments and deleted those which had been demolished. For each local authority area a sample was selected of one in ten of the new hereditaments for which a proposal for valuation had been made since 1 April 1961. Those which were obviously non-residential were deleted from the sample. In addition, each local authority provided a list of the caravan sites containing three or more residential caravans in its area. Finally the Ministry of Defence and the Home Office provided lists of married quarters built since 1 April 1961 for the armed forces and the prison service. One in ten of these dwellings was selected for the census. The addresses of the private dwellings, non-private establishments and caravan sites selected by these methods were written into the 1966 census enumeration record books, initially at the Census Office at the time of preparing the Plans of Division and at a later stage the Census Officers added last minute additions to the rating records which were sent direct from the Valuation Officers. As the sampling method was new a test of the procedures involved was carried out in one enumeration district in each of thirty local authority areas in England and Wales selected at random and in nineteen enumeration districts within a further nine local authority areas purposely selected as being likely to produce special problems of enumeration. Enumerators were supplied with lists of addresses, compiled as proposed for the census. The addresses were visited and, after interviews with house¿holders, were classified in various ways to check the sampling procedure. After examination of the enumerators' records, a re-interview check of about 800 addresses was carried out by members of the staff of the General Register Office. This concentrated on addresses which appeared to have been difficult to enumerate or seemed to have been wrongly classified, in order to obtain first-hand knowledge of problems which would be met in the census. The classified information was used to develop the processing procedures, particularly those concerned with the treatment of sample addresses in buildings containing more than one household. Because of the necessity for tightening up the recognition of buildings, dwellings, households within dwellings and the sample addresses much more detailed instructions were given to enumerators and these were supplemented, for the first time, by formal training sessions given by the Census Officers who had, in turn, attended training given by officers from the General Register Office. The basic intructions to enumerators thus became relatively straight¿forward; they were told to visit only the addresses listed in the record books. If an address was a single private dwelling a census schedule was issued to the head of each household living there. If an address was only part of a house converted to the use of several households, a complete enumeration was made of every household so that it could be ascertained whether the building was, in fact, one or more than one dwelling according to the census definition. Finally, if an address was a 'small' hotel or some other 'small' non-private establishment a special 'Institution' form was delivered to the person in charge. All 'large' non-private establishments were included in the enumeration, but only one in ten of the people present actually completed a schedule except in hotels and boarding houses where everyone present completed a schedule and the sample was later selected at the Census Office. Each caravan site was visited by an enumerator who compiled a list of caravans and issued a schedule to one in ten of these, ignoring holiday caravans most of which were empty at this time of the year. A random number between one and ten was allocated to each site to start the sample. No caravans were enumerated, therefore, on sites where there were fewer caravans than the random number. All the private households attached to non-private establishments were enumerated as such and again one in ten of these were later selected at the Census Office. It will be readily understood that the sampling arrangements entailed the enumeration of rather more than ten per cent of the population and there was a deviation in the number of dwellings in the sample. There were several factors liable to cause the latter circumstance, including Thus the enumerated figures needed adjustment and special procedures were initiated at the Census Office to ensure that subsequent analysis of the data was firmly based on ten per cent of the private dwellings and small non-private establishments and ten per cent of the population in the larger non-private establishments. (For sampling error see 7.2). In 1966 the number of questions asked was slightly more than asked on the 1961 sample 'long' form (E.10/W.10). There were only three topics which were entirely new in 1966 but others required more information than in 1961. The new topics were: For those with a job at any time during the week ended 23 April 1966 the respondent was asked to give information in respect of: Topics on which extra information was required in 1966 included: Three topics were dropped for 1966: nationality (information on birthplace alone being required), fertility, and the use of the Welsh (and Gaelic) language. The processing of the main data for the 1966 Census was, in general, carried out in a very similar manner to that used for 1961 material. However, the Department acquired an IBM 705 computer of its own, which was installed at the Census Office at Titchfield. Possibly the greatest single advantage gained was that by using the same type of computer as in 1961 re-programming was cut to a minimum. The fact that for 1966 only a sample of dwellings was enumerated did not mean that there was a substantial reduction in the amount of published data and as a decision was taken to complete the primary processing in ten months, means were sought to improve the speed of production of material for photolitho reproduction, without too great a decline in quality. Therefore, a combination of computer printed output and pre-printed base sheets was used for most tables although manual typing was used for tables without a standard format. In 1966 a post-enumeration survey was again carried out. Census Officers were asked to take part but not enumerators as the interviewing which they had carried out in 1961 was undertaken by the Social Survey Division of the Central Office of Information. The post-enumeration survey was again taken as soon as possible after the census and there were three main enquiries in England and Wales. In the series of In the No Preliminary Report was produced; nor were separate volumes on Age, Marital Condition or Birthplace; (these data were generally covered in the The provision of statistics for urban wards and rural civil parishes (Ward Library) had been introduced in 1961; similar figures for census enumeration districts were produced only to specific order. In 1966, however, data for all three types of area were produced and made available at a standard charge. This facility proved very popular and these statistics have also been used in aggregate form in the production of tables for Parliamentary Constituencies, Regional Hospital Boards, etc. Also, after the 1961 Census, there had been a number of complaints that in the County Reports the policy of restricting the publication of some tables to certain areas (by type and population size) and restricting the table content for other areas made it impossible to construct statistics for 'ad hoc' areas of interest in planning, particularly at the sub-regional level. Accordingly, it was decided, irrespective of level of publication, to make all tables available for all local authority areas in the form appropriate to the largest type of area (Scale A). Copies of the resulting unpublished tabulations became freely available for the cost of reproduction. The main differences between Scotland and England and Wales in the pre-census testing, sampling procedures, processing of returns and in the post-enumeration survey work are noted below. A pre-test was held in April 1964 which involved about 1700 households in 22 areas of Scotland. The main purpose of the exercise was to evaluate the proposed method of drawing the sample for the census. In Scotland the basic sample was drawn in a different way from that adopted in England and Wales. The main sampling frame was the 1964/1965 Valuation Roll (operative from May 1964) for each of the counties and cities, supplemented by a list of dwellings occupied for the first time between May 1964 and February 1966, which was supplied by local valuation assessors, and by information supplied by town and county clerks about houses built under the auspices of local authorities and coming into occupation during March 1966; from these Rolls etc, Census Headquarters selected on a random basis one in ten of the structurally separate private dwellings. Local authorities also supplied lists of caravan, hut and chalet sites. The addresses of every non-private establishment (hotels, hospitals, etc) were extracted from the Valuation Rolls; it should be noted that there was no attempt to identify 'small' establishments and enumerate these on a sample basis. The arrangements for sampling and enumerating chalets, etc and non-private establishments were similar to those for England and Wales. Special Study areas were selected for 100 per cent enumeration because more detailed information than a sample enumeration would have provided, was needed for special economic and planning studies. To obtain sample data for these areas 'sample addresses' were selected in the same way as for other areas in Scotland. The six Special Study areas which were the subject of a All coding and preparation of punch cards from 1966 data for Scotland was done in Edinburgh; but, as in 1961, computer processing was again carried out on the same machine as that for England and Wales. The joint processing of the 1966 information meant that the data for Scotland were almost identical with those for England and Wales. This can be seen from the number of joint volumes published on a Great Britain basis. The Post-enumeration survey carried out in Scotland was limited to the check on the adequacy of the sample frame and the check on derelict, vacant and occupier absent accommodation; the Social Survey did not take part in these operations. A special survey was, however, mounted in Glasgow by the Social Survey Department to investigate a discrepancy in the distribution of households by rooms which had become apparent during the preparation of the 1966 Census Housing Tables. This survey was subsequently extended to a 1/300 sub-sample throughout Scotland. [i]
The Trend and Pattern of Fertility in Great Britain. A Report on the Family Census of 2946. By D V Glass and E Grebenik. Papers of the Royal Commission on Population, Volume VI. HMSO 1954. [ii]
S.S.H.A., First or Second Scottish National Housing Company in Scotland [iii]
See A quality check on the 1966 Ten Per Cent Sample Census of England and Wales
by Percy Gray and Frances A Gee, London HMSO 1972.PART 2 Significant developments in the scope and organisation of the census
2.1 Background
2.2 List of census dates and reference to relevant legislation, 1801-1966
2.3 The census from 1801 to 1831
The first census, 1801
Houses
Inhabited
By how many families occupied
UninhabitedPersons
Total
Males
FemalesOccupations
Persons chiefly employed in Agriculture
Persons chiefly employed in Trade, Manufacture or Handicraft
All other persons not comprised in the two preceding classes Changes made in the enumerations of 1811, 1821 and 1831
The Parish Register Act, 1812
Changes in Parish Register returns, 1811 to 1831
Rickman's estimates of the 18th Century population of England and Wales
2.4 Changes from 1841 to 1891
The conduct of the census passes to the General Register Office, 1841
The last Parish Register Abstract
Scope of the census extended in 1851
1861 England and Wales
Scotland
Changes introduced, 1871-1891
2.5 First part of the twentieth century
Centenary census of 1901
1911
Changes in procedure
1921
Changes in questions
The fourteenth census, 1931
2.6 National Registration, 1939
Great Britain and Isle of Man
Enumeration procedures
Tabulations
Differences in procedure, etc for Northern Ireland
2.7 The census from 1951 to 1966
1951 England and Wales
Enumeration procedures
Sampling
Changes in questions
Post-enumeration survey
Publications
Unpublished data
Differences in procedure, etc for Scotland
1961 England and Wales
Enumeration procedures
Sampling
Validity of sample
Changes in questions
(1) Education
- professional qualifications held and the main branch of science or technology applicable.
(2) Internal Migration
- details were asked of length of stay at usual address if more than one year or previous address if less than one year. When coupled with other data, information about the amount, direction and characteristics of population movements within the country became available (see also comments under 'Usual Residence' for 1966).
Preparation of the Preliminary Report
Post-enumeration survey
Publications
Unpublished data
Differences in procedure, etc for Scotland
1966
England and Wales Sampling
Pre-test
Instructions and training
Enumeration procedures
Accuracy of the sampling frame
Changes in questions
Processing
Post-enumeration survey
Publications
Unpublished statistics
Differences in procedure, etc for Scotland
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.