London's population and the 2011 Census - London Regional Committee Contents

3  Preparations for the 2011 Census in London

  • The challenges the ONS faces in conducting the 2011 Census could well be exacerbated by the fact that the population, particularly in London, is less compliant and less likely to respond to surveys than it was 10 years ago.
  • Many London boroughs and the London Council for Voluntary Services are concerned that ONS has not allowed sufficient time for adequate preparation to engage with the population about the 2011 Census.
  • Many London boroughs have expressed concern about a lack of engagement with ONS.
  • ONS and its contractors for the Census would benefit from the specific expertise and knowledge of London local authority staff.

ONS preparations for the 2011 Census

Increasing response rates

47. ONS informed us that it was undertaking a range of actions to tackle the risk of a reduced rate of response to the 2011 Census:

    " […] we are putting in place a number of changes to help us address [low response rates]. […] They include, briefly: building a high-quality address register to underpin the whole operation; questionnaire tracking, so that we know which houses have responded and which haven't, on a daily basis; targeting our non-response follow-up effort more intensively and flexibly than before, with three or four times as much follow-up effort in London; new questions focused on getting the count right; an online option to complete for the first time; notwithstanding some of the concerns expressed by local authorities, we have I believe stronger partnership working with local authorities than ever before; and outsourcing some of our activities."[43]

Changes to the questionnaire

48. The Census questionnaire has been expanded to include questions which may help to gather more detailed information about the complexity of London's population.[44] This will include new questions on national identity and citizenship and new questions on the date of entry to the UK and the intended length of stay for in-migrants.

Changes to those counted

49. In London, the Census "will cover everyone usually resident in London on Census night, with a subset of information also collected from visitors present on Census night"; information is also to be collected from "residents in communal establishments and individuals or households with no usual or physical address."[45]

50. In 2001, the Census counted persons at their place of residence whether or not they were actually present there on Census night, a change in practice from previous Census exercises, where persons were counted at the address where they were physically present and the numbers of visitors were transferred back to the area of usual residence.[46] Reviews of the 2001 Census indicated that a failure to attempt to capture visitors, and a lack of clarity about where visitors should be recorded, was a factor in the acknowledged undercoverage of the 2001 exercise.[47]

51. In 2011, therefore, the number of persons ordinarily resident at a dwelling and the number of visitors at that property on Census night will be recorded, though visitors will also be required to complete a census return at the address they are ordinarily resident when they return there, if that date is within six months of Census night.[48]

52. We recognise that the inclusion of short term migrants (those resident for between 3 and 12 months) and those with second homes on the 2011 Census ought to result in future Census based population data which better reflects the demand that these elements of the population make on central and local government services. We welcome the decision explicitly to include short-term migrants and those with second homes in the scope of the 2011 Census. We expect these steps to contribute to greater accuracy in the Census in London.

Work to boost response rates

53. The ONS also assured us that it was working closely with local authorities in an attempt to bolster response rates among "hard to count" populations, and was participating in local authority and community liaison programmes for this purpose. The rationale for this partnership-working was to seek to benefit from the knowledge local authorities have of the communities they serve.[49]

54. The ONS plans to increase the concentration of staff in London in order to identify and then focus on hard-to-count areas.[50] Its hope is that, by directly tackling this issue, the level of response from these areas will increase.

Follow-up on non-responders

55. As in 2001, ONS plans to use the Census Coverage Survey (CCS) to capture data from a 1% random sample of people after the Census has taken place. This data is then used to impute into Census outputs information about those who did not participate in the Census. We heard concerns that, because of the complexity of London's population, the CCS methodology would not robust enough for this purpose, but Census Director Glen Watson was confident that the 2011 CCS would be up to the task:

    "It will be random but stratified to the hardness to count, if you see what I mean, so there will be more effort where it is most needed. [The CCS field force] job is to find every single property, household and address in that area and to try and elicit a response from them. It is rather different. It is a doorstep interview with just a few simple questions. It is not a full Census questionnaire. The results from that are then matched against the Census so that we know who was counted in the Census but not the coverage survey, who was counted in the coverage survey but not the Census, and who was missed in both. We then use some statistical processes to make estimates of the numbers that were missed and that will need imputing…We then have some other processes to try and fill in the rest of the gaps."[51]

56. The ONS are confident that by improving the enumeration rate for the Census in London there would be less need to rely on the CCS:

    "There are two issues, really, in making a good quality population estimate for a local authority. First, it is bad news from a population estimates point of view if we get very significant deviations in response rates between different local authorities. Where Glen [Watson] says that we are focusing on the hard-to-count areas, the difficult areas, it is absolutely the right thing to do in terms of getting the sort of population estimates we need to underpin resource allocation. So that is the first thing: the higher we can get those, the better.

    The second thing we do is the CCS. That is independent of the main Census, but the higher we can get those response rates the better, because it implies that you can do adjustments and reconcile the data better afterwards."[52]

A National Address Register

57. In order to have a more sophisticated understanding of the population which it needs to reach, to support the 2011 Census the ONS is creating an up-to-date National Address Register with information drawn from data held by the Ordnance Survey, the Royal Mail and local authorities. The need for a high-quality and up-to-date address register was highlighted by a number of independent reviews of the 2001 Census by a range of bodies including the Treasury Select Committee, the National Audit Office, the Statistics Commission, the Committee of Public Accounts, and the Local Government Association.[53]

58. Attempts by local authorities to access this list in order to readily identify relevant areas has been hampered by the fact that the individual intellectual property rights in elements of the address list belong to Ordnance Survey and Royal Mail, not ONS. This issue is discussed at greater length below.

Better imputation

59. Significant criticisms were made of the imputation methods used in certain under-enumerated areas in the 2001 Census: the Statistics Commission, which audited the outcome of the 2011 Census, found that the One Number Census methodology used for imputation in 2001 "may have a tendency to work imperfectly in the most hard-to-count areas such as Westminster, and did so in 2001."[54] Glen Watson told us that there had been changes to the imputation methodology following results of the Census Coverage Survey:

    "We have made some changes to the way the coverage adjustment process will work; for example, stratifying the sample so that more effort goes to the areas where it is needed most is new. We also plan to hold back some of the CCS sample, because we acknowledge that we do not yet know—and we won't know until the responses start coming in from the main Census—where it is needed fully. So we are holding some back to allocate it later in real time as the operation is going on. We have also made some changes, which are quite technical, to the way the imputation is actually going to work and the way the gaps are going to be filled in."

Review of outputs before publication

60. ONS told us that the draft outputs of the 2011 Census would be reviewed against other population data and, if necessary, revised before release:

    "The goals for 2011 are to maximise the overall level of quality of data and to minimise the differences in quality between areas. ONS has developed a quality assurance strategy for investigating and, if necessary, adjusting the results before final release of output. This will involve validation with data from other sources to ensure that outputs for areas from the national to the local level, and for specified population sub-groups, have acceptable levels of accuracy."[55]

61. We welcome the commitment of ONS to discuss its proposed outputs from the 2011 Census with local authorities before publication, and to revise them if necessary. We consider this to be a potentially significant step forward in meeting the concerns of London boroughs about the accuracy of Census outputs. Noting the lengthy process which resulted in the adjustment of Westminster's 2001 Census outputs, we recommend that any discussions over the outputs of the 2011 Census in London be resolved speedily and fairly.

Public engagement

62. Census Director Glen Watson told the Committee that he recognised the problems with apathy amongst the public but believes that this could be tackled, and consequently the enumeration rate improved, in 2011 through targeted engagement.

63. In their oral evidence, the ONS spoke of their specific plans for engagement:

    "One example would be the local authority partnership plan that we will put in place with each London borough and local authority in the country. That will map out the detail of how we are going to do this local community engagement. The partnership plan will be constructed by the area managers, when they are in place, who will work very closely with the community advisers. For example, we are going to employ over 50 community advisers, half of whom, I think, will be in London. They will focus on providing dedicated support to the various ethnic groups that we think are going to be most challenging in terms of numbers and getting the response up—for example, Black Caribbean, Indian, Bangladeshi, Chinese and so on. Those community advisers will work with a whole host of local groups on the ground with the area managers and will map out events, communications, talks and community support for the Census.

    A lot of the discussion we have had so far has been operating at a higher national level. We've had many offers of support from different religious groups—for example, to provide access to temples and mosques come Census day or to provide help points and advice centres for people in their communities. There is a lot of good will, and work has gone on nationally to gain traction on the ground. We need this field force out there and we need these community advisers in place. As I say, that will come later this year."[56]

After 2011: better migration statistics

64. ONS also recognises that the production of population estimates requires effective cross-government working and in consequence has launched the Migration Statistics Improvement programme. This includes the use of a greater range of administrative and other data to improve migration and population data. The first stage of this programme is due for completion in May 2010. ONS told us that it had, as part of this exercise, already engaged with various stakeholders, including experts from London.[57]

65. We acknowledge the significant efforts made by the Office for National Statistics to learn from the outcomes of the 2001 Census, both in London and nationwide, and note with approval the changes which the ONS has made to its methodology and its practical arrangements in order to achieve as accurate a Census as possible in 2011.

66. Welcome as these changes are, it is entirely likely that issues not encountered in 2001 will arise during the preparations for, conduct of and aftermath of the 2011 Census, both in London and elsewhere. We expect ONS to have in place systems to recognise and respond to any fresh risks and challenges as they arise.

  • Boroughs which gave evidence to us were not confident that the imputation methods to be used in 2011 would sufficiently reflect the complexity of their populations or the transformations in London communities over the past decade.
  • The London Borough of Newham told us that the area of the borough which had the highest concentration of 'hard to count' characteristics was not included in the Census rehearsal exercise.
  • Witnesses have expressed concern about the validity of data used in mid-year estimates.
  • Many of the boroughs which gave evidence to us remain unconvinced that the ONS has done enough to prevent a repeat in 2011 of the mistakes made in the 2001 Census.

67. Despite the assurances which the ONS gave this Committee about the measures it is taking to avoid the mistakes of 2001, we have heard several concerns from London boroughs about the arrangements for the 2011 Census. Most notably, several boroughs are not confident that the new imputation methods would sufficiently reflect the complexity of their populations, or the transformations which have occurred in London's communities over the past decade. Boroughs have also expressed frustration with a perceived lack of engagement with them on the part of ONS in its redesigning of its imputation methods.[58]

68. The challenges faced by the ONS in conducting the 2011 Census could well be exacerbated by the fact, acknowledged by Census Director Glen Watson, that the population today is less compliant and less likely to respond to surveys than it was 10 years ago.[59] The Demographics User Group, in its evidence to us, also suggested that "the entire statistical landscape is changing" and believed the ONS should respond accordingly.[60]


69. ONS has designated Regional Census Champions "to act as a channel of communication between local authorities and the ONS."[61] London's Census Champion is the Chief Executive of Lewisham Council, Barry Quirk, who is well aware of the particular challenges facing the Census in London. When he appeared before the Committee, he highlighted these concerns:

    "I would certainly like to discuss at some date the methodology in relation to that. I have to say that there are four separate strands of work between local authorities and ONS on this. The first is in relation to the reliability and consistency of the address register in the first place. With that, there's the development control dimension. The second is in relation to electoral registration, because we have lots of experience of that, and the differential response on that as well. We are going to be advising Capita and giving it the information about recruitment. The third is this methodology issue, which is a corporate research and intelligence issue, which feeds into the fourth, which is finance and the distribution of resources subsequently.

    I would say that this is, at the moment, the third-order issue. The first-order issue that we're dealing with is whether we've got the right registers. Are we making sure that we, in local government, are getting the information properly? To what extent are we making sure that all our best efforts are put in place so that we get the best response we can? The methodological issues are obviously something that will come at some stage, and we will have concerns about them."[62]

70. The level of communication between London local authorities and ONS, via the Census Champion or otherwise, remains unclear. Many London boroughs have expressed their concerns to the Committee about a lack of engagement with ONS in its preparations for the Census in London. Barry Quirk himself stressed to us that, as Census Champion for London, he was only one channel of communication among many.

71. In particular, boroughs are concerned that they have not been engaged with by ONS with regard to the address register. For example, boroughs complain that they are not being given sufficient access to the register so that they can check it against locally held data. Where boroughs have brought anomalies in the address register to the attention of the ONS they then hear nothing further about this information. Boroughs believe that if the ONS were to engage with them more fully they could share their wealth of experience, particularly with regard to the annual update of the electoral register.

72. Boroughs have also expressed significant concerns over ONS's willingness to discuss with them the methodology it plans to use for calculating and imputing outputs from the Census. The London Borough of Southwark, for example, would welcome discussion of imputation methodology on the basis of the local knowledge it has about its resident population.

73. The London Borough of Lambeth has suggested that greater efforts should be co-ordinated where boroughs share similar enumeration problems:

    "However, there are several London boroughs with similar enumeration issues to Lambeth, and we would welcome a co-ordinated approach to this from ONS."[63]

We recommend that ONS co-ordinate its approach to tackling issues of under-enumeration across those London boroughs where similar issues have been identified.

Public engagement

74. ONS told us that as early ago as 2006 it had begun to develop its Census Community Liaison Programme for 2011, with specific objectives to "encourage participation in the Census by publicising the Census and explaining its use and value; help provide a potential source of field staff; and provide help and guidance to local communities and individual members of the public."[64] It cited a number of London-based organisations it had already engaged with, such as the Polish Social and Cultural Association, the Federation of Irish Societies and Race On The Agenda (ROTA).[65]

75. Many of the local authorities that the Committee received evidence from were nevertheless concerned that the ONS have not allowed sufficient time to adequately prepare for engaging with people. This indicates to us that the Community Liaison Programme may have had limited impact to date.

76. The London Voluntary Services Council (LVSC) acts as an umbrella group for around 60,000 organisations within London, thereby representing a significant proportion of the city's voluntary workers. We were concerned to hear from LVSC's Director, Peter Lewis, that they had, when they gave evidence to us on 1 February, received no contact from the ONS:

    "Unfortunately we've not heard from it [the ONS] at all and we are not aware of the work it is doing in London, although we haven't had much time to reach out and ask many organisations. Those we have asked have not had any engagement with it as yet."[66]

77. Mr Lewis cited evidence from the Audit Commission that voluntary sector support is essential in connecting public service initiatives with disengaged groups.[67] He doubted whether, given an apparent lack of communication and time, sufficient approaches to disengaged communities could be made to have a positive effect on Census response rates.[68]

78. Happily, LCVS have now been contacted by ONS, as Glen Watson told us on 22 February:

    "I'm very interested in having some discussions with the London Voluntary Service Council and other bodies like it across London. There is still some more that we can do there. A good outcome of this Committee was that I read the evidence that colleagues gave. We have already contacted them; we have already followed up on that first hearing."[69]

79. We welcome the approach made by the Office for National Statistics to the London Voluntary Service Council to seek dialogue on improving participation rates in the Census. We are nevertheless concerned that a programme of community engagement apparently initiated in 2006 had failed to contact a prominent body representing the voluntary sector in London until February 2010. We recommend that ONS take immediate steps to review and intensify its work with the voluntary sector in the run up to March 2011.

80. ONS intends to address concerns over arrangements to verify migrant data by producing a 'local authority partnership plan', outlining details for a stronger engagement between the local authority and community.[70] Boroughs are also planning collaborations with local community and faith groups, in the hope of engaging those communities who have previously recorded a low response rate to the Census[71]. However, community groups have raised concerns about the short window of time available to forge links with disengaged communities.

81. Time to make adequate preparations for meaningful community engagement in the 2011 Census exercise is running worryingly short. We encourage ONS to build on its existing contacts and to accelerate its plans for community engagement through the voluntary sector in the run-up to April 2011.


The October 2009 rehearsal in Newham

82. In the run up to the 2011 Census ONS has carried out a number of practice exercises. Two of these have been in London: a census test was held in Camden in 2007 and a census rehearsal was held in Newham in October 2009.

83. The Newham rehearsal addressed a sample group of 38,850. Although no official figures are available from the rehearsal, Newham council have estimated the response rate at 23%, which the Borough considers to be worryingly low. ONS has unofficially estimated the response rate at 28%.[72]

84. In part, the purpose of this rehearsal was to practice the operation of the 2011 Census's various data processing systems. We heard from Damian Highwood, Westminster City Council's Analysis Manager, that Westminster had been keen to be used as a test area for the Census because of the problems it had faced in 2001. However, ONS did not ask Westminster to participate in the rehearsal, and as a result they remained concerned that in the 2011 Census "the sort of properties that will be counted are those with very simple household structures, and those that are missing will be quite complex and large household structures."[73]

85. We also heard from Michelle von Ahn, Senior Demographic Adviser at the London Borough of Newham, that the parts of the borough selected for the rehearsal were not necessarily those which contained populations which were the hardest to reach and would provide the most rigorous testing of the Census's imputation methods. Indeed, according to Ms von Ahn, having supplied the ONS with maps and details of the whereabouts of Newham's hardest-to reach groups, the borough's statisticians had been surprised that "the area that came out high in all those characteristics was avoided in the Census rehearsal, an area that we thought would have been quite a useful means of focusing on how well the Census could be conducted."[74]

86. Newham believes that the low turnout can be attributed to the failure to use enumerators with local knowledge to carry out the Census. However, ONS suggested that rehearsal turn out is often low, noting that turnout in rehearsals prior to the 2001 Census was around 30%, and that a far better rate of return had been secured in its rehearsal of post-Census coverage surveys:

    "You have to consider that the Newham rehearsal was a voluntary exercise. It's quite usual in census tests and rehearsals to get fairly low response rates, particularly in inner-city areas, when all you can say to people is, "This is a voluntary exercise. We would be grateful if you could help us with it. It is important for helping us get our processes right for 2011." The first question is, "Do I have to?" "No, you don't." "Well, I won't bother then." That's quite a commonly held view, which is confirmed by some of the analysis of reasons for non-response that we did after the Newham rehearsal.

    Interestingly, the final rehearsal response rate for Newham was about 27 or 28%. Across all rehearsal areas, the response rate just tipped 40%. That is lower than previous tests and rehearsals, which echoes the point that I was making right at the start: we acknowledge that the climate is getting difficult. What was slightly more encouraging was that the response rate for the rehearsal of the Census coverage survey, which, as I have said, is very different because it is just a quick knock on the door, a few quick questions and off you go again, was in the high 70s. It was 70% or 72%, I think, in Newham. People will respond either if it is quick and they can do it on the doorstep, or if they have to. I'm quite confident that the low response rates in Newham will be considerably higher come 2011."[75]


The National Address Register

87. We are encouraged by the development of a national address register for the 2011 Census. Such a register is vital for a successful Census in London.

88. ONS believes that the creation of the address register is "progressing well": nationally, 216,000 anomalies in the register have now been identified and passed to local authorities for resolution, and a similar number had been sent to the Royal Mail. 46,000 of the anomalies between the address register and local authority databases, approximately 21%, arose in London boroughs, and as of early March all but four boroughs had examined the anomalies and returned corrections. 51,444 anomalies between the draft register and the Royal Mail address database had been identified in respect of addresses in London boroughs, and the ONS estimates that approximately 80% of these anomalies have also been resolved.

89. The accuracy of an address list in London is difficult to maintain for a number of reasons: there are many houses in multiple occupation, and properties are often subdivided without formal notification being given. During the Census rehearsal in Newham, 1267 new addresses were found.[76] Some London boroughs fear that their local knowledge is not being fully utilised in the preparation of the register, in particular with respect to the construction of new properties. It seems to us odd that the creation of a national address database should not seek to take all available information into account.

90. We note the work which ONS has already done to check and resolve conflicts between the draft register and local authority and Post Office databases. It is nevertheless important not to underestimate the challenges of maintaining an accurate register of addresses for London in the months which remain before Census night. We encourage ONS and London boroughs to maintain and increase present levels of collaboration, in order to ensure that the address register produced for London is as accurate as possible.

91. We understand that the address register will not be maintained in its present form after 2011, despite the substantial time and effort which has gone into establishing and updating it. Shaun Flanagan of the Cabinet Office told us that when ONS negotiated the contract with the Royal Mail, Ordnance Survey and the Local Government Information House to provide data for the register, a condition of the agreement was that the register would not be re-used, but that any improvements to the data would be fed back to the three providers.[77]

92. The Chair of the UK Statistics Authority has already written to Ministers to make the case for the national register to be maintained beyond 2011.[78] The Minister for London told us that negotiations on the future use of the register were ongoing: "there is no dispute about the importance and benefits of resolving this."[79] That view was echoed by Keith Dugmore of the Demographics User Group, who described it as a "golden opportunity to produce a definitive national address register and to keep it going".[80] We were nevertheless disheartened to receive no clear answer from the Government on the issue of lead responsibility for negotiating an agreement.[81]

93. An accurate and well-maintained national address register is an invaluable tool for the 2011 Census, and will be vital for any future exercises to quantify London's population. We find it barely credible that the address register developed for the 2011 Census at substantial effort and expense is to be abandoned following the Census for reasons connected to the ownership of the intellectual property.

94. We concur with the UK Statistics Authority in recommending that the address register prepared for the 2011 Census be maintained as a public resource. We recommend that the Government urgently seek to resolve any outstanding issues with the maintenance of the register after April 2011, and to provide sufficient resources for its continued maintenance and development.

Trust and participation

95. One significant challenge to the goal of reaching a high enumeration rate in the 2011 Census in London is the attitude of the public to participation in the Census. Professor David Martin of the Royal Statistical Society and the ESRC Census Programme told us that

    We see the response rate to official surveys declining by a small amount year on year, which would imply that the public are becoming less willing to take part in official information-gathering activity. Also the research suggests that people think the Government already know lots of this stuff and query why they are asking them.[82]

96. Tony Travers identified trust in the Census process as a key issue which the Office for National Statistics should focus on in the period leading up to Census night:

    "[We should be concentrating on] anything that can be done to promulgate the idea that whoever you are and wherever you live, the more you sign up, the better it is for your area. We should create a sort of civil sense that it is good for everybody, whether it requires celebrities to promote it or whatever."[83]

97. Concerns over confidentiality of data and data-sharing within Government are at the forefront of the public consciousness in a manner perhaps not envisaged a decade ago. Londoners are far more aware of the capacity of Government bodies to share the data they hold on citizens between them. No matter how robust the procedural and legislative safeguards in place to prevent inappropriate and unauthorised data sharing, issues of trust and data security are bound to be raised during the run-up to the Census and on Census night.

98. We well understand the purpose of the Census, the independence from Government of the UK Statistics Authority and the Office for National Statistics and the scrupulous safeguards which are in place to prevent data sharing and the disclosure of personal data. Others, for any number of reasons, may not: some, for example, may fear that accurate completion of a Census return will result in data being passed on to other central and local government departments and agencies, prompting reviews of benefit status or eligibility for council tax discounts.

99. The purpose of the Census, and the use to which the information will be put, may need careful explanation to those resident in London who have fled repressive regimes or who are in communities with no tradition of engagement with the authorities to provide information of this nature. Colin Barrow, Leader of Westminster City Council, suggested that certain migrant groups may, because of their countries of origin, have an understandable inherent suspicion of providing information about themselves and their families to authorities:

    "Then there is fear of authority—a lot of people come here because they don't like where they've been, and it's authority that is the reason why they don't like where they've been; lack of cultural understanding—some countries have no tradition of Census, engaging with authorities and getting involved in all that, and I am briefed that Somalia is one such."[84]

100. A crucial element of the Census exercise is the confidentiality of the personal data collected. We are satisfied that the arrangements ONS has put in place for maintaining the confidentiality of the information gathered in the Census are robust. We recommend that ONS and local authority campaigns urging participation in the Census emphasise not only its confidential nature but also the impact that non-completion or under-completion of Census returns can have on local authority funding and ultimately on local provision of services.

Delivery and collection of Census forms in London

101. ONS has indicated that the majority of Census forms will be posted out to addresses on the address register under contract with Royal Mail. Decisions on whether to post out or to hand-deliver forms will be made will be determined by a number of factors, including

  • confidence in the accuracy of the address list in any given area;
  • the proportion of known, or suspected, multi-occupied properties in the area; and
  • concentrations of large households.[85]

102. In areas which are to receive Census forms through the post, ONS plans delivery to take place in the working week of Monday 14 March, thirteen days before Census day, giving a week as contingency to resolve any issues, and giving time to receive feedback on and to investigate addresses missed and any undelivered questionnaires.[86]

103. In areas selected for hand delivery of forms, delivery will take place between 15 and seven days before Census Day, which ONS says will leave a six-day contingency period for resolving any local delivery problems. Up to three attempts to make contact and deliver the Census questionnaire packs will be made at each household address, after which field staff plan to put the pack through the letterbox if no contact is achieved. Any apparently vacant properties will be recorded.[87]

104. ONS told us that the accurate delivery of Census forms would be underpinned by the new national address register, and that a real-time form-tracking system would assist in following up any non-response.[88] For the first time in a decennial Census, the option of online response to the questionnaire is also available.

105. Professor David Martin explained to us the rationale behind the decision to post out the Census forms to the majority of the country:

    "[…] Fundamentally the 2001 Census seemed to work very well in most parts of the country. What happens, therefore, is that the model of sending an enumerator to knock on the door of a lot of essentially compliant and reasonably easy to enumerate areas is an extremely expensive way of conducting a census. The rationale, which I think is sound, is that sending forms postally to the majority of the country, and needing a relatively light follow-up in order to get a good response, releases much more resource to concentrate on the really difficult areas, in which London is so rich."[89]

106. Census Director Glen Watson told us that the allocation of resources to hand-delivery and follow-up on collection of Census forms in London and elsewhere was being prepared on the basis of models which took account of the experience of 2001 and which drew data from other sources. Although final allocations had not yet been made, he expected that in London, where approximately 12.5% of the addresses to be reached in the Census are, 27% of the 155 area managers for England and Wales would be deployed, and 20% of the 28 to 29,000 Census collectors would be in operation.[90]

107. He indicated some of the methods used by ONS in determining how to allocate resources to checking addresses on the new address register and following up to collect Census forms:

    "For targeting the address check […] we do look, primarily, at where there are mismatches between the different address sources. As it turns out, those mismatches are often clumped around areas with high levels of flat conversions, high levels of students, high intermingled mixes of commercial and residential properties, new developments—that sort of thing. These are important factors in deciding where we're going to do the address checking. They are contributory factors in deciding where we're going to need to put our follow-up effort in chasing non-response."[91]

108. London boroughs and the GLA are increasingly concerned that these final allocations, and the detailed plans for how ONS intends to undertake delivery and collection of forms in London, have not yet been released: these are considered crucial to the success of the Census in hard-to-count areas in the boroughs.

109. The difficulty of making contact with Census respondents in hard-to-count areas in London was repeatedly drawn out in the evidence we received. London has far more homes than any other part of the country where it can be difficult for enumerators to gain access. These include large blocks of flats, houses of multiple occupation, and properties which have only recently been constructed and therefore may not feature on maps. John Hollis of the GLA told us that in his view hand delivery of Census forms had to be concentrated on areas of London with high numbers of hard-to-reach properties, such as houses in multiple occupation.[92]

110. Colin Barrow, Leader of Westminster Council, explained how certain residences can be difficult to access, and how this difficulty can be exacerbated when the residents are also not used to dealing with the authorities collecting personal information:

    "It isn't door knocking at all; it's entry phone pressing. If you visualise for a moment what it is like when you are in a three-bedroom semi and someone knocks on the door, the immediate instinct is to open it. But if you are in a block of flats in a difficult environment and there is a buzzer with somebody with some official-sounding gobbledygook on the phone and you do not understand it, because it is not your first language and the entry phone does not work very well and crackles a lot, you hang up and that's it—the person does not get any further. They do not get further with the guy next door either. Some 85% of our properties in Westminster are like that."[93]

111. Councillor Tim McNally, Southwark Council's executive member for resources, told us that

    The vast majority of our borough is difficult to count. To relate it to electoral registration, for example, when we post it out, we get a 25% response. When we go out on foot, we get up to 90%. That is the difference. The posting route just doesn't work for us in Southwark.[94]

112. Eileen Howes of the Greater London Authority highlighted the lack of information from ONS on the resource it planned to devote to hard-to-count areas of London:

    "[…] At this stage, we do not know how [ONS] will focus on the hard-to-count areas. We are concerned about the local people being employed. We hear that 5% of areas will be hand-delivered, and then we hear that it might be lower. We are left in the situation where we do not know how they will deal with the hard-to-count areas."[95]

113. Professor Martin, of the ESRC Census Study Programme, was concerned that, although the preparation of address lists was a dynamic process, ONS was now leaving it late to determine the detailed plan for delivery of Census forms:

    "[…] What we would really love to see now is the detailed plan for which areas are being targeted for hand delivery. That may be—should be—the thing that makes it work … I am now of the opinion that this is getting late. If you asked that question even a few months ago, I would have said no. There is no point checking address lists and doing detailed planning for something that you know is a very dynamic process. It has got to be done quite late to be right. The lists exist—it is not that ONS do not have the lists—but the question is doing the checking and getting it back in the time scale for full delivery."[96]

114. Keith Dugmore of the Demographics User Group was similarly keen to see the delivery plan: "my feeling would be that a great deal of effort and budget ought to be put in London to try to get the response rates above 90%."[97]

115. Glen Watson recognised the significant logistical difficulties which faced the Census in London boroughs in actually making contact with residents: "I don't disagree that it will be very challenging … the biggest difficulty is the initial contact."[98] He was nevertheless cautious on committing too much of the resource allocated to the 2011 Census to London:

    " … I have a limited budget. I am responsible for conducting a Census in the whole country, not only in London, so I have to do a good job everywhere. I don't think it would be good use of money to put all of the resource into London and nothing anywhere else to try to get the London response rates up to the levels that we find elsewhere. There might be a point where you'd get diminishing returns. By the time you've knocked on a door or tried to remind a household for the 10th or 12th time, you're fighting a losing battle. That is why some of the other mechanisms that we have in place are important."[99]

116. The level of resource which ONS is prepared to allocate to hand-delivery and collection of Census forms in London boroughs is one key to the success of the Census in London. We recognise the balance which must be struck between ensuring a high response rate to the Census in London and ensuring that the Census operation for England and Wales as a whole is properly resourced. It is nevertheless crucial that, given the challenges facing the Census in London, the proportion of Census staff allocated to enumeration in London is adequate to the task.

117. We are concerned to note that ONS has not yet published its detailed plans for areas which have been selected for hand-delivery of Census forms. These plans should be a principal element in the strategy to deliver a successful census in London, but their value will be diminished if they are not shared with other stakeholders—such as London boroughs and, indeed, London Members of Parliament—who have a detailed knowledge of the areas concerned and are aware of the difficulties encountered in getting to the front doors of certain properties. We recommend that ONS publish for discussion, as soon as is practicable, its detailed plans for hand-delivery of Census forms in London, and engage London boroughs and other stakeholders in dialogue about the most effective ways to ensure that Census forms can be delivered and collected in hard-to-count areas of the capital.

Recruitment and training of Census staff

118. ONS plans to employ some 35,000 temporary staff across England and Wales to support the 2011 Census. The contract to recruit and train Census staff was awarded to Capita in March 2009.

119. Many witnesses who gave us evidence expressed concern over the decision to provide centralised recruitment and training, fearing that those recruited centrally would not receive training adapted for the area that they were enumerating, and would not have sufficient knowledge of localities where they would be expected to deliver and collect forms.

120. Representatives of the boroughs who gave oral evidence to us on 8 March had been keen to encourage ONS to second local authority staff to work with the Census field force in the run-up to Census night, but had not received a positive response from ONS: "ONS has given a list of reasons why it would be too difficult to put into its arrangements the secondment of local authority staff to the Census enumeration process."[100]

121. Glen Watson was confident that by running a recruitment campaign across England and Wales, recruitment of people in each locality would present no difficulties:

    "Capita recruits the people for us even though they are ONS employees. Our contract with Capita makes it clear that we expect the ethnic mix of the workforce to reflect the ethnic population from which the workforce is drawn. Expectations are written into the contract, and quality impact assessments will be done to make sure that the processes will deliver what we want."[101]

    On the local recruitment side, I have read and heard a lot of concerns about the local recruitment in Newham. All I would say is, come 2011, we will be recruiting everywhere. We will not be faced with the problems of having a select number of jobs advertised with people from all around London interested in applying, because we will be advertising jobs everywhere, so, naturally, there will be a much stronger local element.[102]

122. Eileen Howes of the GLA, who had worked in the field for the last three Censuses, told us she had been monitoring the Capita website advertising Census jobs which are currently available, for instance in checking the address register, to track the London boroughs where jobs had been advertised. She was not impressed:

    "If you look at the Capita website advertising these jobs, most of east London was listed under the East of England region, not under London at all: Ilford, Barking, Dagenham, parts of Newham—not all of it. I had previous enumerators from 2001 coming round looking for where the jobs in Ilford, which is where I worked, were advertised. They were advertised under the East of England region. When they phoned the website, they were told Ilford was with Uttlesford in Essex. Actually, it's not; it's a different place called Redbridge. It's a complete disaster in terms of employing local people because local people in east London want to apply and they can't find the jobs. It doesn't bode well for employment of collectors and enumerators.[103]

    [Capita] said they would look at it. Some of east London was then moved back but not all of it. Then a lot of Berkshire appeared in London. So Reading is in London … You will never get local people employed [in London] if the jobs are advertised under the East of England region, because people will not look there. That is an example of things that are likely to go wrong if you don't employ local people."[104]

123. In his evidence to us, Professor David Martin argued that rather than local geographical knowledge, the most important tool for a successful Census was cultural understanding:

    "I don't think [local knowledge] is essential, and this is very much a personal view of the process, rather than something on which there is a big body of research. It is not essential that somebody knows the neighbourhood in the sense of knowing the streets, because the work being done on the preparing, mapping and address-checking in a sense provides a uniform information base to enumerators. However, it is probably very important that the enumerators are people who are acceptable to the community being enumerated.

    If the enumerator is going to be doing door-to-door visits, they should be someone who understands the community on whose doors they will be knocking. That is not a sense of knowing which street is round the corner, or where the back flat is, but is much more to do with understanding what languages they may encounter and what kind of cultural objections to the Census there may be in that local neighbourhood, because those with that understanding are much more likely to be successful at the door. That is borne out by all the large surveys and the way in which survey field forces are deployed."[105]

124. Eileen Howes also indicated that, from her experience in a previous Census, cultural understanding was crucial to achieving good enumeration results:

    "We had particular streets where there were largely Muslim families and they would only open the door to someone who was or who looked like a Muslim. We just picked the right enumerator and sent them down and they got all the forms back. There was no problem. It was a fear of opening the door to people who did not look like them or looked like white men in suits, for example. You need the flexibility in the street in London to send the right people round. We are missing out on that at the moment."[106]

125. Census Director Glen Watson believed that the training to be given to the Census field force would be adequate to enable enumerators to deal with several of the challenges to be encountered in London:

    "Training packages of e-learning, DVD and classroom-based training will pull together the field force and take it through the steps and processes. They will include something about the barriers and difficulties that might be expected in the areas. I have done that training. It is very good. It includes role play for dealing with someone who cannot speak the language. It includes role play for dealing with entryphone systems that are just buzzers we cannot get past, people refusing and people not being bothered. It is a simulated environment. It is role play. It is quite effective."[107]

126. We nevertheless know from our constituencies in London that those enumerators chosen to go door-to-door in London boroughs will encounter individuals and families from a remarkable variety of communities and backgrounds, more than are likely to be encountered in a standard training exercise. No one enumerator will be able to be trained and equipped to deal with each of the languages and cultures which might be encountered across any neighbourhood assigned for enumeration. Duncan Whitfield of Southwark Council and Colin Barrow of Westminster Council both highlighted the difficulties which may arise with languages on the doorstep:

    "We are looking particularly at some very deprived areas and the difficult places to go, where even with door knocking, it is a challenging arena to go into … when politicians do go door knocking [in Southwark], they find some properties where the only people who speak English are the children.[108]

    [In Westminster] only 30% of children in school speak English as a first language, so that means 70% of their parents probably didn't, and very possibly still don't … "[109]

127. Barry Quirk believed that ONS and its contractors could benefit from the specific experience and knowledge of London local authority staff:

    "So the important issue is the way in which Capita recruits and the extent to which local authorities' electoral register people, who have managed the electoral register, say from late September to late November, can use the lists of people that they have to advise Capita, so that when Capita is hiring it is actually hiring people who know the area and do work like this in the area anyway. I think it is crucial that people know the area rather than just being brought in, because they simply won't know the local turf otherwise."[110]

We concur. We recommend that in its final preparation for the Census ONS should work closely with local authorities to learn from their experiences of electoral registration.

128. We consider that diversity in the 2011 Census field force is going to be crucial to ensuring high response rates in areas which have been selected for hand enumeration. We are concerned to note that, despite the assurances given about the Capita contract, the recruitment process for the 2011 Census field force in London may not be geared to provide a sufficiently diverse pool of enumerators.

129. We recommend that ONS and Capita work intensively with Jobcentre Plus and local jobs partnerships in London boroughs to ensure that advertisements for 2011 Census field force roles are actively targeted to reach as diverse a cross-section of London's communities as possible.

130. Following recruitment, we recommend that ONS takes particular care in assigning enumerators to neighbourhoods where forms will be delivered and collected, to ensure the best match between enumerator and local community.

43   Q 125 [Glen Watson] Back

44   Draft questionnaire pages have now been published in the Census Regulations 2010 (SI., 2010, No. 532). Back

45   Ev 54 [Office for National Statistics] Back

46   Cm 7513, para 3.1 Back

47   Cm 7513, para 3.2 Back

48   Cm 7513, para 3.6 Back

49   Ibid  Back

50   Q 130 Back

51   Q 134  Back

52   Q 135 [Guy Goodwin] Back

53   Ev 56 [Office for National Statistics] Back

54   Statistics Commission, Report No. 22, Census and population estimates and The 2001 Census in Westminster: Final Report Back

55   Ev 57 [Office for National Statistics] Back

56   Q 149 Back

57   Ev 60 [Office for National Statistics] Back

58   Q 57  Back

59   Q 125 Back

60   Ev 81 [Demographics User Group] Back

61   Q 178 Back

62   Q 137 Back

63   Ev 72 [London Borough of Lambeth] Back

64   Ev 58 [Office for National Statistics] Back

65   Ibid. Back

66   Q 38 Back

67   Q 39 Back

68   Q 46 Back

69   Q 182 Back

70   Qq 66 and 149 Back

71   For example, the London Borough of Hackney [Ev 79] Back

72   Q 139 [Glen Watson] Back

73   Q57 [Damian Highwood] Back

74   Ibid  Back

75   Q 139 Back

76   Q80 [Sir Robin Wales] Back

77   Q 207 Back

78   Letter of 8 July 2009 from Sir Michael Scholar KCB to Rt Hon John Healey MP, Minister of State for Housing, Department for Communities and Local Government, available at . Dr Tony Wright MP, Chairman of the Public Administration Select Committee, and Councillor Ian Swithenbank CBE, Chairman of the LGA Group's Improvement and Development Agency for Local Government, have also written to the Minister in similar terms.  Back

79   Q 208 Back

80   Q 30 Back

81   Shaun Flanagan (Q 211) suggested that the Department for Communities and Local Government had lead responsibility for determining the issue. The Government's Advisory Panel on Public Sector Information (APPSI), in its submission to DCLG's consultation on policy options for geographic review from Ordnance Survey, has indicated that lead responsibility for resolving the issue has now passed to the Department for Environment, Food and Rural Affairs: response dated 9 March 2010 on . Back

82   Q 35 Back

83   Q 36 Back

84   Q 74  Back

85   Helping to shape tomorrow, Cm 7315, para 4.20. Back

86   Ibid. para 4.21 Back

87   Ibid. para 4.22 Back

88   Ibid. Back

89   Q 24 Back

90   Q 130 Back

91   Q 146 Back

92   Q 119 Back

93   Q 94  Back

94   Q 83 Back

95   Q 119 Back

96   Qq 27-28 Back

97   Q 31 Back

98   Q 151 Back

99   Q 152 Back

100   Q 96 [Damian Highwood] Back

101   Q 110  Back

102   Q 182 Back

103   Q 107 Back

104   Q 108 Back

105   Q 7 Back

106   Q 108 Back

107   Q 168 Back

108   Q 94 [Duncan Whitfield] Back

109   Q 74 [Colin Barrow] Back

110   Q 169 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 2010
Prepared 31 March 2010