Select Committee on Treasury Minutes of Evidence

Memorandum by Mr Len Cook, Registrar General for England and Wales and National Statistician


  1.  Censuses were held in all parts of the United Kingdom on Sunday 29 April 2001. Following devolution responsibility for the 2001 Census in Scotland and Northern Ireland was transferred to the Scottish Parliament and Northern Ireland Assembly, respectively, and the appropriate Registrars General. This Memorandum covers the position in England and Wales, although some references are made to the parallel arrangements in Scotland and Northern Ireland.

  2.  The Census is not only the central core of the system of population and labour force statistics on the UK, but it is also central to statistics on health, environment, education, housing, ethnicity, family and communities. It underpins the population projections and estimates for later years that are vital for Government and business planning and are widely used within both the public and private sectors.

  3.  The Census has been described as the UK's biggest peacetime exercise, which requires the goodwill of the public. It is a large, intensive operation carried out in a very short time frame everywhere, regardless of difficulty. But, with a response rate which is estimated to be 98 per cent (see below), the 2001 Census is expected to meet its objective to deliver the information that is required for the decision-making process both nationally and locally. Central and local government, health authorities and other organisations will use the resulting data in planning housing, education, health and transport services for years to come.

  4.  Forms were delivered to households in the three weeks running up to Census day. For the first time the public were asked to post them back, and only households that did not mail forms back were visited again by enumerators. The law requires every household to complete and return a form, and a large-scale publicity campaign was mounted to raise public awareness.

  5.  Scanning, image recognition and coding of the forms is underway at a purpose-built centre in Widnes, and further processing and analysis of the data will begin shortly at Census headquarters at Titchfield in Hampshire. The first outputs will be published in August 2002.


  6.  This year marks the 200th anniversary of the Census in Great Britain. During these two centuries—with the exception of World War II—a Census has been conducted at least once every 10 years.

  7.  As society has become more complex, and demand for information has become greater, the number of questions on the form has increased. In modern times, a primary objective has been to balance the needs of data users against the form-filling burden on members of the public, while ensuring each Census delivers maximum value for money.

  8.  In planning the 2001 Census, the Office for National Statistics (ONS), set out to learn as much as possible from the experience of Census offices around the globe. There was regular liaison with Census-taking organisations throughout the developed English-speaking world. INSEE (France), other European countries through Eurostat (European Statistical Office) were also consulted, and still other organisations through meetings of the UN where the ONS played a major role in developing the UN guidance for the 2001 Census round.

  9.  Information was exchanged on such topics as the use of geographical information systems, the development of scanning and recognition technology, coverage measurement, the use of the public mail service for the return of Census forms, fast tabulation output systems, community liaison, disclosure control and the protection of confidentiality. As part of this process, ONS helped produce Diffusion, which is an occasional international newsletter aimed at providing a forum for the exchange of experiences for Census planners across the English-speaking world.


  10.  The UK Government's proposals for the 2001 Census were set out in a White Paper (CM 4253) published in March 1999. They were based on four broad strategic aims:

    —  to ensure that the question content is appropriate to meet the demonstrated requirements of users;

    —  to deliver products and services to meet legal obligations and users' needs within stated quality standards and to a pre-defined timetable;

    —  to ensure that all aspects of the Census data collection operation and the dissemination of results are acceptable to the public and comply with Data Protection law; and

    —  to demonstrate that the Census represents value for money.

  11.  The previous Census was based on broadly similar principles, but, in a number of respects, the 2001 Census took a different approach to cope with difficulties encountered in the 1991 Census. These difficulties arose from changes in society, technological developments, and the need to satisfy Government requirements for information to support policy initiatives on, for example, social exclusion, inequalities in health, crime prevention, carers and the new deal for the unemployed.

  12.  The total budget for the 2001 Census in the UK over the 13 year period 1993-2006 is £254 million. The breakdown for the period 1993-98 and subsequent years is:
(£ millions)
1993-981998-991999-2000 2000-012001-022002-03 2003-042004-052005-06 Total
19.612.421.4 53.7107.020.1 254.1

  £207 million of these costs relates to England and Wales.

  13.  The largest elements of the total cost are to pay for the delivery and collection of the forms and for the processing of the data.


  14.  The topics on the Census form are those that have been shown to be most needed by central and local government, the health service, academics, businesses and professional organisations. In each case, no other comparable and accessible source of the information is available in combination with other items in the Census. Consideration was given to the public acceptability of topics and to whether or not questions could be asked in a way that would elicit reliable answers. The cost of processing the answers to questions was assessed in relation to the usefulness of the results. Finally, the overall length and layout of the Census form was considered so that the burden on the public was kept to an acceptable level within the overall objective of achieving optimum value from the Census.

  15.  The topics were designed to be mutually supporting; that is, each one should provide information that will make others more useful. This is a particularly valuable aspect of a Census, where information on a range of topics is collected simultaneously for the whole population to form a single source from which important inter-relationships between two or more topics can be analysed. Answers from the individual people forming households and families will thus be able to be combined to provide information on the number and characteristics of households and families of different types, such as, for example, the number of single parent families where the parent is employed and the children are under school age.

  16.  A copy of the England household form is attached at Annex A. [1]


  17.  The physical conduct of the Census is a huge logistical operation with a field staff in England and Wales of some 69,000 and a target audience of every household in the country, Each enumerator was responsible for the delivery of a Census form to each household in their district. To help them achieve this, they were equipped not just with a map, but also an address list detailing every known address. They were trained to make contact with as many households as possible (for it is known that this helps people understand the purpose of the Census and thus the response rates), but also to note new addresses that might not be on the address book in their records. In Wales they were trained to offer each and every household a form in Welsh and a form in English and the option of completing either.

  18.  Communal establishments such as hospitals, prisons and residences for the elderly were treated differently and each person was required to complete an individual form. There were no exceptions and special arrangements were made, for example to enumerate the Armed Forces.

  19.  Approximately 30 million Census forms were delivered either by this method or by posting them out to those households who had inadvertently been missed. Once completed in respect of 29 April 2001, the forms were returned via the Royal Mail to the 2000 Census District Managers who were required to log returns and arrange for personal visits to those who had not completed their Census form or who were having difficulty in so doing. The returned forms were collected from each of the Census District Managers and transported to the central processing centre in Widnes, where they will be processed over the forthcoming 10 months in the custom built environment using scanning, recognition and auto coding technology. The Widnes site is staffed by 1,200 people specifically trained for the purpose.

  20.  Following the Census, a Census Coverage Survey took place. This survey was designed to measure how the Census did in counting households and people by interviewing a cross section of the population and carefully matching the results from both the survey and the Census. In England and Wales the survey covered approximately 300,000 households in 20,000 postcodes selected to form a representative sample. By using the Census and Census Coverage survey in combination it will be possible by August 2002 to provide a full and complete count of the population.

  21.  This approach, often referred to as the "One Number Census" methodology is designed to estimate for missing items of data, and to allow for missing people. The methodological development was undertaken by a joint ONS-academic team. Professor Ian Diamond of Southampton University, a world expert in this field, was in charge of the academic input. The work was overseen by an expert steering committee, and the system was examined in detail both by ONS's specialist methodologists and by independent academic experts. The quality-assurance process for this methodology included a detailed presentation to the Royal Statistical Society.

  22.  The final stage of the production of outputs includes quality assurance of the figures against other information, including demographic analysis and statistics derived from administrative data such as child benefit or pension records.

  23.  The Census timetable is attached at Annex B.


  24.  The Census collects information from each person and household in the country. But it is not concerned with facts about individuals as such. Its purpose is to provide facts about the community, and groups within the community, as a whole. The public has a right to expect that information provided in confidence will be respected.

  25.  Precautions will be taken so that published tabulations and abstracts of statistical data do not reveal any information about identifiable individuals or households. Special precautions may apply, particularly to statistical output for small areas.

  26.  Independent reviews of security and statistical confidentiality were carried out and the recommendations of these reviews were acted upon. A report on these reviews was laid before Parliament in advance of the start of Census operations.


  27.  The Census is a devolved matter, and separate Censuses are held in Scotland, where the General Register Office Scotland (GROS) is responsible, and in Northern Ireland, where the Northern Ireland Statistics and Research Agency (NISRA) runs the operation. The ONS is the responsible department in England and Wales. In Wales, both as part of the normal process of consultation, and in the context of the debate over the lack of a "Welsh" tickbox in answer to the ethnicity question, ONS worked closely with both the Wales \Office and the National Assembly for Wales. In the run-up to the Census, over a period of years, ONS consulted with Welsh user groups and the Welsh Language Board.

  28.  At the UK level, the objective was to adopt common approaches wherever appropriate but recognising the different circumstances and requirements in each country. ONS, GROS and NISRA worked together to ensure that differences in logistics and form content were as few as possible, consistent with the views of the different Administrations. Senior management from the three Census offices and the Welsh Office (since devolution of the National Assembly for Wales) came together to oversee the planning process on the UK Census Committee. The three Census offices also are represented on the Census Programme Board, responsible for the day-to-day management of the Census, and on other committees reporting to that board.

  29.  Across England and Wales, 101 area managers were recruited, and in Wales an overseeing senior manager was appointed. These senior officials were expected to take responsibility for running the Census, to oversee the recruitment and training of the entire field force, and to deal with logistical problems at area level. They also played a key role in the publicity campaign. The recruitment process was designed to ensure the successful candidates had the necessary skills, and the Census area managers (CAMs) then were given two full weeks of training.

  30.  The training covered such topics as Census strategy and philosophy; likely problems and issues; lessons to be learned from the 1991 Census; team working; ethnic and disability issues; and dealing with the media.

  31.  The first task of the CAMs was to recruit the district managers—The next tier of the field staff and absolutely key to the success of the Census. Then came the team leaders and the enumerators, with the latter responsible for delivering the Census form to householders, and for following up where necessary. Training was cascaded down the organisation, with specially-shot videos forming an important element in the training process.

  32.  The 69,000 field staff in England and Wales included 62,500 enumerators, compared with a field force of 115,000 in 1991. The reduction was made possible by the decision to use the postal service for the return of the completed forms.


  33.  The primary legislation that provides for the taking of a Census in Great Britain is the Census Act 1920; in Northern Ireland the corresponding legislation is the Census Act (Northern Ireland) 1969. Under the current terms of these Acts, Orders in Council, which may prescribe:

    —  the date on which the Census is to be taken;

    —  the persons by whom and with respect to whom the Census returns are to be made; and

    —  the particulars to be stated in the returns,

  are required to be approved by both Houses of Parliament.

  34.  The Schedule to the Census Act 1920 authorises the inclusion, in the Censuses for Great Britain, of the following matters in respect of which particulars may be required:

    —  names, sex, age;

    —  occupation, profession, trade or employment;

    —  nationality, birthplace, race, language;

    —  place of abode and character of dwelling;

    —  condition as to marriage, relation to head of family, issue born in marriage; and

    —  any other matters with respect to which it is desirable to obtain statistical information with a view to ascertaining the social or civil condition of the population.

  35.  Under the terms of the Census Act 1920, the Register General for Scotland is responsible for taking the Census in Scotland. Following devolution it fell to the Scottish Parliament to approve separate subordinate legislation relating to the specific arrangements for the Census in Scotland.

  36.  The Registrar General for Northern Ireland is similarly responsible for making arrangements for taking the Census in Northern Ireland. Separate Northern Ireland subordinate legislation as provided for in the Census Act (Northern Ireland) 1969 was introduced and approved.

  37.  Under the terms of the Census Act 1920, the Registrar General for England and Wales has responsibility for conducting the Census in Wales. Authority for conducting the Census in Wales was not, under the terms of the Order in Council made under the Government in Wales Act 1998, to be devolved to the National Assembly for Wales. Thus the same procedures applied for making an Order in Council and Regulations jointly applicable to the Census in both England and Wales, as in previous Censuses. The Registrar General for England and Wales recognised, however, the importance of gaining the support of the Welsh Assembly for the arrangements for the 2001 Census in Wales and sought to ensure that the Welsh Assembly was fully informed of all such arrangements.

  38.  Following approval of an Order in Council the Chancellor of the Exchequer laid before Parliament Census Regulations which made detailed provisions for the conduct of the Census. These included specimens of the forms to be used.

  39.  The proposal to include a question on religion in England and Wales in the 2001 Census required a change to the primary legislation, since the Schedule to Census Act 1920 does not, as currently worded, permit such a question to be asked. Such an amendment was necessary before a question on religion could be specified in the subsequent Order in Council for England and Wales.

  40.  For a brief history of the Census Parliamentary process, see Annex C.


  41.  Consultation on the Census started in 1995. Three working groups were set up to consider the content of the Census (questions to be asked), population definitions and output. The first two were key to shaping the Census proposals, upon which more formal consultation was carried out with ONS's established Census Advisory Groups. There are five advisory groups in England and Wales, with one for each customer sector—central government, local government, health sector, academics, and business sector—and one in Scotland and one in Northern Ireland.

  42.  In addition, the Census offices welcomed views and submissions form any source, with or without invitation, and information papers were issued regularly.

  43.  Proposals for the Census process were tested extensively, with more than 40 small-scale question tests. There was also major test in 1997 to evaluate new collection and processing methods including post back of the Census forms, as well as alternative styles of Census form, and to test public reaction to some new and revised questions—notably a question on income. As a result several of the procedures were revised substantially, and at a later date the income question was dropped.

  44.  In 1999 there was a Census rehearsal to test the delivery and collection procedures (including post back), the public acceptability of the form, the follow-up coverage survey, and the systems for processing the data. The rehearsal covered nearly 150,000 households throughout the UK, in areas chosen to include a cross-section of population and housing types.

  45.  As a result several of the procedures were revised substantially, and at a later date the income question was dropped. In addition, neither the test nor the rehearsal raised any significant issues in Wales, other than the need to issue a separate Welsh language form to all households. Following debate in Parliament the religion question, unlike all the other questions, was not compulsory.


  46.  As part of the drive to improve the cost-effectiveness of the Census operation, options for outsourcing were considered. A study in 1994 defined a strategy and made detailed recommendations. The costs and risks were then analysed by the Census Offices. The recommendation that the contracting out of the administration of the payroll was viable was accepted and tested in 1997. Following the successful trial a contractor was appointed to operate the payroll for the 1999 Census rehearsal and the 2001 Census itself. The 1997 test also enabled the concept of using automatic data capture and coding technology to be trialled. This involved scanning and image recognition techniques as well as automatic and computer-assisted coding of write-in responses. The test proved the techniques to be viable and contracts (which included the printing of forms) for the rehearsal and Census were awarded following an open options procurement process in December 1998. Other areas contracted to the private sector include the printing of other Census supplies, the delivery and secure collecting of all Census materials to field staff, the operation of the public helpline and publicity.

  47.  The advantages of contracting out as well as delivering efficiency gains, were seen to be a reduced need to divert ONS management resources into the logistical task of setting up, maintaining and running down a facility designed for a one-off task. It also enabled ONS to access technology and other expertise outside ONS's own core expertise. A detailed evaluation of all aspects of the Census operation is being carried out. Inevitably, some of the contracts have given rise to more issues than others. Aspects of the Census operation have proved a challenge to the contractors concerned and to ONS staff with responsibility for the Census. Legitimate concerns have been raised about the payroll and helpline which need to be evaluated. At the time of writing, Census processing is taking longer than originally scheduled, although the Census Offices, with the contractors, are invoking contingency options in order to meet the deadlines for output delivery within the overall Census budget.

  48.  All in all, in terms of innovation, this year marks a watershed for Censuses in the UK. As well as contracting out, the 2001 Census has also seen the first use of the post for the return of the forms, and the extensive use of tickbox answers which has made it possible to increase the number of questions without increasing the load on householders completing the Census form.

  49.  The cost of the 1991 Census in England and Wales was £117 million. To compare properly with the 2001 operation it is necessary to allow for a 10 per cent increase in the number of households and for the increased difficulty in contacting households. Additional tasks recommended by the National Audit Office to improve the planning of the Census and the requirement to code 100 per cent of responses to all questions (in 1991 only 10 per cent of a number of questions were processed) also added to the cost. After further allowing for inflation over the period, the business case for the Census estimated that on a like for like and consistent pricing basis the cost of the 1991 Census would have been £263 million against £207 million for the 2001 Census, an overall efficiency gain of 25 per cent.


  50.  It is presently estimated that the response rate will reach 98 per cent of households; with around 88 per cent of the estimated number of households returning forms by post, a further 7 per cent being collected by hand by the enumerators and of the remainder, possibly 3 per cent, being vacant dwellings and second homes. While we will not know the precise response rate for individuals until all the forms are processed, we expect it to be as good, if not better, than achieved in 1991 (when the response rate for individuals was 97.8 per cent).

  51.  With such a response rate in exceptionally difficult circumstances (such as the foot and mouth outbreak) the Registrar General is satisfied that overall the 2001 Census will be shown to be a success.

  52.  The Registrar General will consider taking action to prosecute for non-compliance in those cases which have been reported to him and where there is evidence of a clear refusal to complete and return a Census form. So far 86 cases have been referred to the Solicitors' Office, but, for a variety of reasons, not all of these will go to Court. A number of cases are now reaching the Courts resulting in some prosecutions.

  53.  Although voluntary, the Census Coverage Survey scored a response rate of over 90 per cent, which augurs well for the accuracy and completeness of the "One Number Census" process.

  54.  Only when the Census and the Coverage Survey have been completely analysed will ONS have the final and definitive picture of how successful the Census has been. The size of the Coverage Survey is sufficient to generate coverage adjustments for each local authority, and will make possible the creation of a complete dataset for analysis of individual communities at a small-area level. This will be a marked improvement on 1991.


Foot and mouth disease

  55.  The foot and mouth epidemic proved an unexpected complication in conducting both the Census and the coverage survey. However, special arrangements were invoked following discussions with MAFF and the Welsh authorities, as well as the farming unions. The Census was already using post back and therefore enumerators would not need to visit premises as often as previously, thus the procedures were more straightforward to adapt.

  56.  A co-ordination unit and dedicated database were established, and intelligence exchanged regularly with the National Farmers Union. This enabled up-to-date information to be supplied to the field staff, the Census helpline and the National Statistics website. In affected areas, Census forms were delivered by mail and to make this possible additional material (20 different items) was printed and distributed at short notice.

  57.  Enumerators in rural areas were given special training to enhance written instructions. All Census staff were ordered to comply with all signs forbidding or restricting access, not to stray from paths or metalled roads, to comply with disinfecting arrangements, and if in doubt to post out forms. In addition, an incident database was established to facilitate speedy resolution of problems. In the event, the Census passed off without serious incident, and evidence to date suggests data quality will not suffer in the countryside.


  58.  The Census helpline was contracted out to the private sector. Neither ONS nor the contractor envisaged the volume of calls received, and the service initially was overwhelmed. Many callers were frustrated at the length of time they were kept waiting, their inability to get through to an operator and the number of times they had to call before getting satisfaction. ONS responded by opening hundreds of extra lines and arranging a separate service for members of the public requesting forms. All in all, the number of lines was increased from 350 to 1,300 within a matter of days. Over the three-month period the helpline was operational, more than 2.6 million callers accessed the automated interactive voice response (IVR) system designed to provide answers to the most commonly asked questions, and of these nearly 404,000 spoke to an operator. In addition, more than 12,000 responses were made to e-mails received from members of the public visiting the National Statistics website.

  59.  From the beginning, and throughout the planning and operational phases, ONS staff worked closely with the contractors to ensure that the helpline was responsive to public needs. A specially developed database, maintained by Census staff, was used as a central repository for a range of operational information and briefing material. Information on the database ranged from guidance on completing each of the questions on the Census form to detailed material on the development and testing of selected questions—including those on ethnicity and religion.

  60.  During the Census, senior staff held daily briefing meetings to identify issues of current concern, based on an analysis of calls to the helpline, media coverage, and feedback from Census staff dealing with press and public. In this way, it was possible to update the briefing material available to helpline operators so as to meet changing requirements, and to deal with issues of special concern. Particular account was taken of the daily report of the questions most frequently asked by callers to the helpline. For example, at different times additional material was prepared on the foot and mouth epidemic, the recording of second homes or holiday accommodation, the campaign for Braille Census forms, and non-compliance procedures.

  61.  New material was transmitted and disseminated as widely as possible using the IVR system, the helpline operators and the National Statistics website. In addition, the material was fed to BBC Local Radio stations across the country, for use in daily Census slots in their programming.


  62.  From July 2000 onwards a campaign was mounted in Wales calling for a Welsh tick box within the ethnicity question on the Census form.

  63.  ONS undertook the extensive consultation in advance of the publication of the White Paper setting out the plans for the 2001 Census, and the 1999 Census Rehearsal included parts of Wales to test not just the questions, but also the procedures and particularly the Welsh and England language issues. The plans included various measures designed to ensure that there was a "Welsh Census" in Wales. These included the appointment of a Welsh-speaking Census manager, the delivery to every household of forms in both Welsh and English, an improved question about the Welsh language, a country-of-birth question with a "Wales" tickbox option, and a dedicated advertising campaign in Wales.

  64.  By the time the Welsh "tick box" issue was first raised, the content of the Census forms had been debated and passed through the appropriate Parliamentary procedures and the logistical exercise of printing Census forms had already begun. There followed a sustained campaign by individuals and organisations in Wales—notably the Western Mail newspaper and the Independent Wales party—with continuing and often vociferous criticism of ONS.

  65.  Within the constraints upon him, the Registrar General sought to respond positively and introduced a package of measures to improve information about Welsh identity and promote the option for those in Wales who wished to record themselves as Welsh to use the "write in" option available on the form. In addition, the Registrar General held meetings in north, mid and south Wales and made several other bridge-building visits to encourage the people of Wales to complete the Census form.

  66.  In the event, the tickbox campaign does not appear to have had any significant impact on the response rate, or on the quality of the responses in Wales. Throughout the enumeration, Wales provided some of the best-responding areas.

  67.  The previous Economic Secretary to the Treasury, Melanie Johnson, agreed that, in the light of the strength of feeling expressed in Wales on the matter of recording Welsh identity in the Census, the National Assembly for Wales should have a more formal role in agreeing future Census forms in Wales. ONS and Assembly officials are currently discussing how this might best be established.

Enumeration in hard to count areas

  68.  There were small pockets where particular difficulties were encountered and where, consequently, lower response rates have been achieved. But ONS was always aware that Census-taking would be tough in parts of inner-London, for example, and planned appropriate measures.

  69.  In an effort to improve the enumeration in these hard to count areas, field-staff workloads generally were half the size of those in less-difficult parts of the country. The assumed postal response in London was 60 per cent, with a ten-day follow-up period to chase the remainder. In the event, the postal response was 80 per cent, and enumerators had to chase only 20 per cent of London households. Logistical difficulties, including the retrieval and analysis of forms posted back, were, however, in many areas more severe than anticipated and enumerator contracts were extended to ensure the count would be as full and complete as possible.

  70.  ONS also was grateful for assistance given by a number of London boroughs, and by voluntary and community organisations across the capital. This was an important strand in efforts to improve response, and undoubtedly has helped in reducing the differential undercoverage across different population groups that had occurred in 1991.

  71.  There also is evidence to suggest that the inclusive nature of the Census publicity campaign, with its "Count me in!" slogan and events aimed at hard-to-count groups—for example members of the ethnic minorities—was successful in convincing these groups that they should complete a form.

Blind and partially-sighted people

  72.  There had been extensive and prolonged consultation with the Royal National Institute for the Blind about the availability of explanatory material in Braille and audio-tape form. ONS was grateful for the support given by the RNIB in helping to prepare these materials. Subsequently, however, RNIB did not consider that the steps taken were sufficient and mounted a campaign in the weeks immediately before the Census for a Braille Census form and other ways for responses to be provided independently by blind and partially sighted people. ONS sought to respond in a positive way within the constraints of the systems that were by then already in place by arranging for blind and partially sighted people to complete forms by e-mail and on the telephone. More should be done for the next Census and ONS will start planning for this before the next questionnaire is developed.

Payment of Field Staff

  73.  Although the overwhelming majority of field staff were paid on time, a significant number were not. This was due to problems with procedures during the intensive field operation and in the period immediately after the Census. Among the issues encountered were poorly completed forms, failure to complete or submit claim forms on time, and errors in the scanning and payment process. In general it has to be said that the procedures for handling the payment of field staff were inadequate. ONS accepts that it did not live up to its own standards for the payment of staff who had worked so hard to make the Census a success. Particular attention must be paid to this aspect in any future Census.

  74.  Once the scale of the problem became apparent, ONS put in place contingency measures to ensure staff were paid as quickly as possible. A dedicated call-line team was created to handle incoming pay queries; expert advice was taken in refining and improving workflow procedures; internal technological expertise was harnessed; and additional staff were trained to handle the queries internally.

  75.  Field staff were advised of the position through letters from the Registrar General.

Incomplete forms

  76.  There was some media coverage suggesting that enumerators were only seeking answers to three or four basic questions on the Census form. This was incorrect. Full information was required from every member of each household, but as a fallback position enumerators were instructed to obtain the basic details rather than obtaining nothing at all. In practice, in dealing with some respondents, this meant obtaining the answers to the three key demographic questions, sometimes by observation.


  77.  Unplanned issues emerged during the Census that had to be addressed in real time and had a substantial effect on the budget profile. In particular, there was the impact of the Welsh tick box, foot and mouth disease, the initial inadequacy of the helpline, and the additional work required—particularly in London—as a result of logistical difficulties in the retrieval and analysis of forms posted back. ONS remains confident, however, that the overall cost of the Census will be within the original £207 million budget allocated.


  78.  Census is protected under the 1991 Census Confidentiality Act and the Census (Confidentiality) (Northern Ireland) Order 1991 and is taken very seriously by the Census Office.

  79.  The physical security of the Census forms, from their completion through to their destruction, was independently reviewed by a specialist security consultant. His recommendations were fully implemented. Statistical confidentiality and the protection of individual information is assured through the disclosure control programme which itself was independently reviewed by a statistical expert form Statistics Canada. Again, his recommendations were acted upon.

  80.  In addition, field staff were required to comply with the requirements of Census legislation, and there was a comprehensive training programme.

  81.  Despite these measures, a serious breach of confidentiality occurred when around 190 completed and partially completed forms relating to some 60 households were among material found by a member of the public in a waste sack outside an East London community centre used by field staff. The Registrar General has issued a public apology, written to all those whose forms were involved, and set up an independent investigation.


  82.  The effort and expense of taking a Census are worthwhile only when the results meet needs, and are delivered effectively. With this in mind, data users have been consulted over a period of years, and a number of innovations will be made to the output-production process to improve the products and services available.

  83.  A major innovation is that national and local data will be released concurrently. Another is that virtually all the products will be electronic, with easy-to-use products available free-of-charge at the point of access, and extensive metadata readily available to help users make effective use of the results. There will be a set of standard products—a pre-defined set of tables of Census counts for all levels of Census geography—coupled with a fast and inexpensive service to respond to the needs of individual customers.

  84.  An extensive consultation process continues with users to tailor the various products and services.

  85.  The publication schedule is as follows:

  August 2002: Mid-year population estimates for 2001 based on the results of the Census. These will provide data broken down by age and sex for each local authority area. The only difference between these numbers and the Census figures will be demographic change between 29 April 2001 and 30 June 2001.

  Early December 2002: Key statistics for areas throughout England and Wales, with all data made available on the National Statistics website. These summary statistics will include the key counts and denominators as well as such indicators as the percentage unemployed in an area and the percentage of the population with long-term illness. Similar products will be produced in Scotland and Northern Ireland.

  First half of 2003: Main national and local results delivered, again via the web and other electronic media, and to DTLR for local authority standard spending assessment (SSA) purposes.


  86.  Systematic and continuous review has formed a fundamental part of the Census programme. Each stage of the process of planning and implementation has been subject to rigorous analysis and evaluation, and the post-Census day phase of this evaluation is already underway.

  87.  The intention is to produce a General Report evaluating the Census, as was done for the 1991 Census. However, the individual chapters will be published on the National Statistics website as they become available. In due course, the complete report, including an overview of the various components, will be published in hard-copy form.

  88.  In addition to the General Report, ONS will publish a detailed Quality Report which will bring together all available information on data quality. Both reports are scheduled for completion by June 2003.

  89.  Together with other review material—some of it already in the public domain—the General Report is intended to cover every aspect of the 2001 Census. These other documents include:

    —  1997 Census Test: evaluation summary report;

    —  1999 Census Rehearsal: evaluation overview;

    —  Census Quality Survey (1999) report;

    —  Security and confidentiality reviews: summary as laid before Parliament;

    —  White Paper: The 2001 Census of Population; and

    —  Census Information Papers.

  90.  As was noted earlier, the 2001 Census marked a watershed and, in planning for the future, a primary requirement is to build on progress made. For example, the community liaison programme and the publicity campaign were undoubted successes, but there are serious lessons to be learned from events surrounding the helpline, the payroll and the Census in Wales. We believe our field management processes can be simplified through more extensive use of available technology. In this and other areas, the Census will be reshaped considerably by advances in technology.

  91.  The use of the internet to file Census returns is only one issue that will require serious consideration.

  92.  As recommended in the report of the Social Exclusion Unit's Policy Action Team on Better Information published last year, ONS is currently taking forward an evaluation of the case for conducting a Census in England and Wales in 2006. This will be published in January 2002 and will provide a basis for further discussion on the way forward.

21 September 2001

1   Not printed. Back

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2001
Prepared 26 November 2001