Project Objective
To put in place the policies, products and services to implement the Census output strategy. That strategy was based on key criteria about user needs and Office for National Statistics (ONS) provision of outputs.
Following completion of the policies and the specification of product and service, many aspects of the production of the output were dealt with by a partner project, Output Production. This is the subject of a separate evaluation report, which includes evaluation of the timetable for release.
Background
The 2001 Census was the first which based the planning of output on a strategy which was agreed at the outset and included in the White Paper (as government policy). It was based on the simple premise that 'the investment in the Census was only justified if the results are made available speedily and in a clear and useable form'.
The Census output strategy comprised four major groups of outputs:
a set of statistical tables ("Standard Area Statistics") which would fulfil user requirements;
products, including Reports to Parliament, which would meet statutory requirements for output and allow easy access to standard census results;
arrangements for licensed and value-added use of 2001 Census output; and
a specialist service to deal with user enquiries and requests for commissioned output.
The key criteria to meet users needs, were that outputs should:
be comprehensive and standard output be comparable between areas;
have flexibility to meet ad hoc requirements;
be fit for purpose and delivered to a pre-announced timetable; and furthermore
their use should not be constrained by bureaucratic or financial barriers, particularly charges.
The key criteria for ONS were that outputs should:
be drawn from a database which was based on 100 per cent counts and as unbiased as possible, with an integral framework of geography and accompanied by all relevant supporting information;
be produced speedily and efficiently and be user-friendly, accessible and intuitive;
minimise the risk of disclosing information about identifiable individuals; and
use definitions and classifications that are harmonised, as far as possible, with other sources.
Methodology
The methodology adopted throughout the project recognised both the need to accord with statutory requirements and Government and ONS specific policies and the importance of finding out user requirements and reflecting these in the product specifications.
Consultation with users
Substantial consultation was undertaken from 1999 to 2003 to ensure that the Standard Area Statistics met, as far as possible, user requirements. Discussions were held with key users to ensure their needs were fully understood. Four rounds of general consultation were undertaken, during which draft tables were made available to all users for comment, and initial plans for specification and release of the output were discussed at Census Roadshow events in venues around England and Wales. The draft tables were amended substantially as a result of the consultation.
Additional consultation was conducted with users in Wales to ensure that results on the Welsh language were presented in the most useful form.
With regard to provision of standard products, consultation also revealed the variety of user requirements in methods of accessing products, from printed volumes to the non-expert users accessing results through the Neighbourhood Statistics Service (NeSS), to developers of commercial packages loading results in bulk into their databases. This led to decisions to make Census results available in a variety of ways.
Setting up 'approved suppliers' list
In developing the charging policy for results from the 2001 Census the team took into consideration the terms of the Census Act, government policy and the ONS policies on release of results and on cost-recovery. With the removal of the requirement for commercial licences, the need for value-added resellers to become Census Agencies disappeared. To help promote high standards in the re-supply of Census results, an Approved Suppliers scheme was created. The Scheme allows organisations to apply for Approved Supplier status - that is, confirmation that the Census organisations have examined the Supplier's arrangements for dealing with census data and are satisfied that they are of adequate quality and that the Supplier demonstrates that it has the required technical and statistical competence.
Specific requirements for Wales
The need for bilingual presentation of census results was reflected throughout the output planning process and comment invited from the National Assembly for Wales; Welsh Language Board; and other Welsh users on these plans.
Provision of information to users
Users were kept updated with the constantly developing plans for releasing the Census results through the National Statistics website, the Census News publication (also available on the website) and a series of roadshows.
Dealing with user inquiries
Central Enquiry Unit was, and still is, able to deal with simple questions on the Census results. In addition, a specialised Census Customer Services team deals with more detailed questions requiring greater expertise or experience of using census data. This team received full training on the 2001 Census output.
UK co-ordination
Plans were co-ordinated, where possible, with other UK Census organisations through Joint Output Project Board meetings. Whilst each census organsation collected slightly different information, and faced different requirements for data from their users, this co-ordination allowed a 'core' of consistent and comparable tables to be produced by each organisation.
Assessment and lessons learnt
The project was successful in meeting the requirements of the Output Strategy. In particular, the extended consultation on content of the output, and the immediacy and extent of information provided to users, undoubtedly improved the quality of the Census products, and the plans for use of those products. Some examples of key achievements are:
Specification of Standard Area Statistics Tables
Consultation with users showed that there remained a preference for most census output to be provided in a standard set of tables, rather than relying on a method of obtaining customised output, though this question should be revisited for output from future censuses. The Standard Area Statistics were specified following extended consultation with users. The tables fulfilled stated requirements for data for the Standard Spending Assessments for both England and Wales as well as reflecting, as far as possible, the stated requirements of those responding to the consultation.
Users kept informed of plans for output
Comprehensive information on plans for output was contained in the Census Output Prospectus on the National Statistics website and further developments in the plans were announced in Census News, also on the website. Emails were sent to users to alert them to any major developments being announced. In addition, two sets of events were held, in Autumn 2001 and Spring 2003, to discuss and explain the plans for releasing the Census results.
Availability of commissioned output
As in previous censuses, it was recognised that standard output cannot meet all requirements for analysis of the Census results. A system has been set up to process requests for non-standard 'commissioned' output.
Dissemination of results with minimal barriers to use
Results from the 2001 Census were disseminated in a variety of media and formats. Cost and bureaucratic barriers to use were also minimised.
Common UK approach
Whilst each Census organisation has its own specific requirements for output, a common approach to table design and numbering was achieved to make use of UK-wide results as simple as possible, in particular, by allowing a set of comparable tables to be produced by each Census organisation.
Integration of Welsh language output
The importance of providing bilingual reports for the first releases of census results, and for detailed results on the Welsh language, was recognised and integrated into the planning of the release of census results.
Lessons learnt
Although the project was successful overall, several aspects could have been improved:
Late changes to specifications were sufficiently substantial that they resulted in changes to the planned output timetable and the numbering and layout of tables, with consequent difficulties for users. A more efficient approach may have been to have a 'mop-up' product to included any additional requirements and to be produced towards the end of the Standard Area Statistics timetable, leaving pre-announced products to be released as planned.
Calls made to Local Authorities following the release of the National Report revealed both positive and negative reactions to provision of the results in SuperTABLE format, with several users disappointed at the work involved in obtaining all results for a single area.
Whilst changes made through the consultation process greatly improved the usefulness of the Standard Area Statistics, the management of these changes could have been streamlined. The tabulation system treated each table as a separate entity so changes affecting a standard classification or single footnote had to be replicated many times, with a consequent risk of error. An approach considering each table as a combination of standard building blocks would have greatly helped to ensure consistency and accuracy in development and reduce manual checking at the output production stage.
The Standard Area Statistics had to be developed before important decisions were made on some aspects of ONS statistical policy, in particular regarding disclosure control and geographic boundaries. In future more should be done to integrate development of statistical policy and development of census products, and ensure that any subsequent changes in policy are considered with an appreciation of user requirements.
There was the inevitable tension between concentrating on the specification and development of the standard Census products and responding to requests for non-standard tables required for other projects. The work on standard products was frequently overridden by the ad hoc requests. One solution is to delay all such requests until after the great bulk of the standard output had been produced.
The need for a specialist customer service team with expertise in the Census output products, was clear from discussions with users, and the number of queries received immediately following the release of the early reports provided further evidence of this need.
Problems were experienced in filling vacancies for relatively short-term (less than a year) jobs to specify the 2001 products. Consideration should be given to how these problems could be avoided in future. The inclusion of more statistical expertise within the Output Policy team might help and might have the added benefit of avoiding some of the problems with integration of policy and development, as noted above.
Conclusions
The project can be considered a success by the criteria outlined above. The emphasis on consultation and communicating with users improved the quality of census products and services. On the basis of the outputs produced so far, the new methods of disseminating the Census results appear to have succeeded in fulfilling user requirements. The final months of dissemination, and any analyses users subsequently undertake, will provide further feedback.
Census News
