Select Committee on Treasury Appendices to the Minutes of Evidence



Memorandum by Dr Ludi Simpson, Centre for Census and Survey Research, University of Manchester


  1.  The census is achieved with a deserved high reputation for consultation, preparation and detailed output that represent extremely good value for the essential data that anchor a decade's public planning. There is no alternative approach that can match the census for democratic provision of high quality data for local areas. In 2001, the weakness of parts of the census fieldwork that were subcontracted for the first time, suggests that this aspect of its organisation can be improved before the next census.


  2.  The census offices' thorough and documented consultation and research won general acceptance for their plans ahead of the 2001 census. Their plans to measure the extent of undercount of residents in 2001, in order to include an allowance for it in all output (the "One Number Census"), are innovative and have been widely and generally welcomed as an appropriate response to the census undercount encountered in 1991.


  3.  The census' social, demographic and employment data for small areas are absolutely essential to the functioning of government. The census is a "numbering of the people" that is independent of all administrative registers. This independent enumeration is the only means of gaining coverage of all people, common application to every small area, a range of basic social and employment topics, and availability of results both inside and outside government.


  4.  It is impossible to review fully the conduct of the census until its processing is complete: only then will the quality of data be measured in full. However, the voices of the temporary workforce who carried out the enumeration in 2001 make up an essential ingredient for reviewing the conduct of the census.

  5.  A compilation of interviews with fieldworkers and a commentary on them have been recently published. They are the only evaluation of the fieldwork so far published and allow firm conclusions to be reached on some aspects. They are used to inform the comments below (Bradford Council published "Collated fieldwork experiences from the UK 2001 Census" in October 2001, and the commentary is contained in "Census fieldwork—the bedrock for a decade of social analysis", Cathie Marsh Centre for Census and Survey Research, Occasional Paper 22, University of Manchester, M13 9PL. The commentary is attached as an Annex to this evidence).[1]

  6.  These interviews with fieldworkers, and ONS documents, highlight some concerns about the general organisation of the fieldwork. The following comments focus on one area, the success of the parts of the operation which were for the first time contracted out as commercial operations:

    (a)  The speed of postal collection of census forms, contracted to Royal Mail, was completely inadequate to meet the timetable demanded by the census operation. The result was delayed fieldwork, and consequently a lack of focus on those households hardest to count, higher undercount, poorer quality results, and considerable additional costs.

    (b)  The arrangements to pay the 62,500 enumerators in England and Wales, contracted to ADP Chessington and partly sub-contracted to LASON, were over-bureaucratic and payments to many thousands of enumerators were delayed. The delays were a matter of days for most, more than five weeks for many thousands, and more than two months for some hundreds. The result was extreme hardship for some enumerators who depended on the income from the work, loss of good-will from the managers who had to deal with the enumerators' demands without any means of satisfying them, and extra costs due to ONS' eventual administration of the delayed payments in-house.

    (c)  The census telephone help-line, contracted to Cable and Wireless, was insufficient to deal with demand for it. This resulted in many thousands of frustrating and costly calls by a public concerned to participate. Census District Managers were concerned that they received many requests from help-line operators unaware of the census aims and procedures.

  7.  These experiences do not suggest that Census expenditure was too great but that the commercial contracts were insufficiently specified to meet fully the needs of the census. A fully specified contract may be more expensive, but achieve the needs of the census operation without the additional costs that were incurred directly by ONS in 2001 due to a service that did not meet the needs of the operation.

  8.  Questions that might be usefully answered in this respect are:

    —  What was the level of service contracted with the Royal Mail for return of census forms to District Managers?

    —  Did the contract stipulate a return of forms within a fixed number of days from collection?

    —  Did the contract contain key dates to recognise that District Managers had to check all forms before telling enumerators where to return? For example all forms posted in the week of the Census, by 4 May, were needed by Managers by May 7 so that enumerators could return to the street on 9 May as instructed by ONS.

    —  What were the different levels of service offered for postal return, and what led to the choice of service? How was the contracted level of service to be ensured?

    —  Similarly, what led to the inability of the pay and helpline services to respond to the census' needs? Which improvements could be made to such contracts in the future?

    —  What progress has the scanning and coding of census forms made to date, compared with that planned for, by Lockheed Martin? What means does ONS have to ensure that there are no delays? Delays to this contract would have a serious impact on the utility of the Census.

    —  What is the proportion of items missing and invalid on forms already processed, compared with the proportion in 1991?

  9.  If the census operation continues to be based around a short time period, as most believe it must, then commercial contracts may not be suited to the jobs they were asked to perform for the 2001 Census. Commercial contracts were not responsive to the behaviour of other organisations and of the public in a timely manner and were not focused on the key aims of the census. The census fieldwork operation needs a more highly co-ordinated management than was achieved in 2001.

  10.  My evidence is written from experience as

    (a)  the responsible user for Bradford Council of 1971,1981, and 1991 census data,

    (b)  the lead local authority officer for consultation on population statistics with government, 1992-2000,

    (c)  principal research investigator at the University of Manchester into the size, nature and implications of undercount by the 1991 Census,

    (d)  external member of the ONS Project Board 1996-2001, planning methods for reducing the impact of undercount on users of the 2001 census,

    (e)  interviewer of census enumerators and their managers during April-July 2001.

31 October 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 2002
Prepared 6 March 2002