User Consultation: Advisory Groups Meetings - November 2002
Summary of issues and topics covered at the round of Census Advisory Group meetings held in November 2002. The groups involved - the Departmental Working Group, Central and Local Government Information Partnership: Census Sub-Group, Health Service Advisory Group, Academic Advisory Group, and Business Advisory Group - cover central government, local government, the health service, the academic community, and the business sector respectively.
Census 2001 update First results
There had been widespread press coverage of the launch of the first results on 30 September, and on the subsequent concerns expressed in the press and by a number of local authorities on the accuracy of the figures. In response ONS had been taking a robust line in defence of the census and in its confidence in the One Number Census process. Bilateral meetings had been held with Manchester and Westminster councils to discuss their particular concerns. Local authorities were being invited to visit Titchfield to see how the ONC process was carried through.
At an RSS meeting on 1 November a QA information pack had been made available illustrating the process that had been followed to assess the accuracy of the census figures for each local authority. ONS were developing a package of material which would also include details of CCS response rates and a dependency paper which it was hoped would satisfy all the local authorities needs. The material was subsequently placed on the ONS website on 23 December [http://www.statistics.gov.uk/census2001/whats_new.asp].
ONS confirmed that revised mid-year population estimates 1991-2000 were scheduled for release on 13 February, and that revisions back to mid-1982 would follow around mid- March. The 1991 Census figures themselves were not being affected. Priority was being given to reviewing the methodology for estimating the migration component of the mid-year estimates, and ONS reported that a National Statistics Quality Review of International Migration had been initiated.
Although there were still some uncertainties in providing firm release dates because of the need to consider further factors relating to the harmonisation of boundaries and overlapping geographies that would affect the timings, a revised schedule had been announced to help users prepare for the receipt of the data. It was now proposed to make available the Standard Tables at a single release for each main geographical level rather than in separate tranches, as had been previously proposed. Availability on the website and via other media would be synchronised.
Output Area (OA) boundaries would fit to those ward boundaries effectively in force as at 31 December 2002, leading to completion of OA production around mid-January 2003. There were still some outstanding issues to be resolved with OS before any boundaries could be released to users prior to the output. ONS would, however, endeavour to make these boundaries available to users at the earliest possible opportunity.
NB: Subsequent decision made that some new ward areas which were due to have an Order signed in mid-January will have OAs planned such that they will nest into the new ward boundaries, enabling full standard tables to be produced for those wards as constituted as near as possible to the production of the data itself, subject to the usual ward population/household threshold limits.
ONS reported that consequence of adopting small cell adjustment was the restriction of exact-fit statistics to a single hierarchical geography (OAs through to LAs). Other geographies would be built from a best fit of wards or OAs to avoid disclosure risks from differencing. Further consideration was being given, however, to the risk of releasing the full Standard Tables for Health Areas, National Parks and Parliamentary Constituencies, which may not always be built from whole wards. ONS would advise Departments when final decisions on these geographies had been made.
NB: Subsequent decision made to release key statistics for National Parks on 13 February on an exact-fit basis.
Output for Health Areas in England, planned for release in May/June would cover the new health area structure. Subsequent release of output for the old health authorities, on a best-fit basis, would be considered after the completion of the main results.
ONS reported that the All Fields Postcode Directory was planned to be available in April 2003 to coincide with the release of Key Statistics for Output Areas.
Current expectations were that Special Migration and Workplace Statistics would become available shortly after the completion date of September 2003 for other products noted in the Annex to the paper.
NB: Subsequent announcement by GROS was that origin-destination statistics were being scheduled for release on 30 September 2003.
The paper updated progress on the specification of the files to supply output to those users not wishing to use SuperTABLE. ONS were developing a programme to convert SuperTABLE files to a CSV format that would be compatible with other dissemination packages such as SASPAC. Further enhancement and testing was being undertaken of the prototype CD using SuperTABLE that had been shown at previous Roadshow meetings.
Advisory Group members were asked to indicate the means by which they would prefer to access output in bulk rather than via NeSS. ONS could provide no facility whereby particular users could be provided with advance release of outputs.
It had not been possible to circulate paper in time for the meetings. However, it was reported that, in due course, the paper would give an overview of the wide range of material that was currently on the website supporting the first results. In addition the paper would summarise those metadata products, such as the Definitions Volume, the Quality Report and the General Report, that would be available both in stand-alone versions, as well as being integrated with the output data itself. Before the paper was finalised it would be necessary to be able to confirm that links to such support material would be equally available to users accessing census output through the Neighbourhood Statistics portal. NB: Paper was subsequently circulated 28 January 2003.
Paper AG(02) 10 had been circulated, paper AAG (02) 02 circulated to the Academic Group only. The Multi-Source Topic Reports would replace the traditional Census Topic Reports that had previously comprised the more detailed cross-tabulations not covered in the local statistics output. Additional reports on 'Religion' and 'Children and Young People' were now being planned.
High-level summary reports on a UK basis would be issued in Summer 2003, followed by more detailed, Focus-style, reports towards the end of 2003/early 2004 both in hard copy and web-based formats.
The paper had been circulated. ONS reported that there had, in addition, been a meeting at the RSS in October. In summary, the paper had stressed the emphasis being given to the importance of protecting confidentiality in the assessment of the proposals to expand the SARs. ONS were now proposing a 1 per cent household SAR similar to the 1991 SAR, and two individual SARs at different levels of geography - plus a possible national file - but constrained by a total sample of 3 per cent.
ONS noted users concerns that it should take a balanced view in assessing real and perceived risk but reported that the strategic position on disclosure control was not likely to change. There was, however, further potential for creating a small area microdata file of limited information with a sample of 5 per cent. Comments had been invited by 29 November.
The paper had been circulated and was supplemented with a PowerPoint demonstration summarising the main components, timetable and interfaces of the programme to consider the longer-term need for, and if necessary, the options for, the future production of census-type information.
The suggestion to use the ACACIA programme (to establish a linked national address list) as a potentially useful interface would be pursued within ONS.
Members' attention was drawn to the proposal to broaden the scope of the Advisory Groups to focus more on the wider needs for small area social statistics. Some concerns were expressed about any unnecessary overlap with the remit of other user groups such as the NeSS domain groups. In the meantime Advisory Group members would be invited to attend a workshop in March 2003, as part of a wider programme of consultation, to discuss the 2001 Census evaluations and to consider the requirement for census-type information and small area statistics, and to discuss the pros and cons of the alternative options for collecting such data, drawing on international experience.
ONS noted that contingency planning for a traditional-type census in 2011 would continue, and that a major consultation exercise on the viable options would be carried out in the summer. Recommendations for future development based the outcome of this exercise and on a cost-benefit analysis of the alternative options, would be presented around October 2003.
Users welcomed the feasibility study into creating a Population Register. Its potential use, as well as the associated confidentiality concerns, was generally appreciated.
Some users regretted the decision not to have a mid-term census in 2006. In the light of concerns recently expressed about the quality of the 2001 Census in particular areas, they regarded this decision as, perhaps, premature.
Strategic planning of the census in Scotland would be developed in parallel.
Paper had been circulated giving a broad background to the Neighbourhood Statistics programme. The most recent data set added was House Prices on 7 November. The Annex to the paper set out those data sets that were currently under consideration.
Paper had been circulated. The valuable contribution that the Advisory Groups had made in the planning of the 2001 Census was acknowledged and it was hoped that a similar level of consultation from a users' perspective would assist in shaping the future of NeSS.
ONS reported that 1998 boundaries had been chosen as the most suitable benchmark year for NeSS and that this boundary set would continue into 2003. A prototype to transfer 1998 datasets into a future compatible format was expected in March 2003. OAs from the 2001 Census would be used a future building bricks, but the geographic hierarchy and the use of so-called 'Super OAs' was not yet decided. The sticking point seemed to be that there was no clear definition of what constitutes a 'neighbourhood'.
The current Terms of Reference for the Advisory Groups already provided for consideration of ways in which other statistics might be integrated with census output to produce improved local information throughout the intercensal period. ONS felt that it was within the remit of the Group to consider NeSS and other relevant issues. However, ONS would consider the views of all Advisory Groups before making any further proposals to expand the scope and membership of the Groups.
Particular issues that Group members wished to see covered included work on small area population estimates and income modelling.
Paper had been circulated for information in August, reporting on a consultation exercise that had been conducted earlier in the year. ONS agreed to follow up the suggestion that there should be some feed back on the results to those who had participated.
Paper had been circulated to Group members in September to ensure that no user with a potential interest in the revised classification was unaware of the consultation. A number of users had concerns about backward comparability.
Date of next round of meetings
In the light of the decision to arrange the Strategic Development Workshop for Advisory Group members in March 2003, ONS proposed that the next round of meetings should be convened around late April/early May 2003.
Census News
