Census (Hansard, 3 May 2001)

The information requested falls within the responsibility of the National Statistician, who has been asked to reply.

Letter to Lord Lipsey from the National Statistician and Registrar General for England and Wales, Office fir National Statistics, Mr Len Cook, dated 3 May 2001.

As National Statistician and Registrar General for England and Wales I have been asked to reply to your recent Questions asking:

(1) what information is available from the census which is not available from other sources; (HL 1845)
(2) what cross-checks are carried out to identify the accuracy or otherwise of the census; (HL1846)
(3) what trials of the census questionnaire were carried out prior to its general issue; (HL1847)
(4) what would be the approximate cost of a sample survey or 1:10,000 carried out by (a) a random telephone and (b) a household questionnaire; (HL1848)
(5) what are the advantages and disadvantages of a full census compared with a sample survey of the population; (HL1849 and
(6) what plans are there to evaluate the cost-effectiveness of the 2001 Census after its completion (HL1850).

106WA

I will answer these in order.

Availability of data

Detailed information on some topics, including relationships, ethnicity, religion, knowledge of Welsh, general health, carers occupation and travel to work is not available from other sources.

On other topics, while it may be the case that some individual particulars of personal information that are collected in the Census may be available from a number of disparate administrative sources, the value of the Census lies in the fact that only in the Census are the many different particulars recorded in combination with each other in order to measure inter-relationships between variables. Furthermore the Census is the only source of data which can provide such statistical information on a uniform basis about the country as a whole and about individual small areas and sub-groups of the population at a common reference point in time.

It is the Census information on basic demographic characteristics that provides the benchmark for population estimates, social surveys and many public and private information sources about government. Thus the Census remains a unique source of information for resource allocation and a wide range of other purposes within central and local government, business, academia, communities and elsewhere.

Checks

Field checks are made at the time of the collection of the Census returns to ensure that acceptable forms are returned. Subsequent coverage checks are made to assess the extent and nature of any under-enumeration and to provide information by which the census counts themselves may be improved. In addition, aggregated census counts at particular geographic area levels are compared with other sources of statistical data for evaluation purposes.

Trials of census questions

The questions on the 2001 Census form were subjected to an extensive programme of cognitive research and small- and large-scale public testing. This included two major census tests in 1997 and 1999 in a sample of areas, selected to reflect a cross-section of various population sub groups. The 1999 test served, in fact, as a rehearsal for the Census itself.

Questions are only included in the Census where a clear need has been identified that cannot be met elsewhere and a successful business case has been made. There must also be clear evidence that they are publicly acceptable and would not adversely effect the response to the Census as a whole, and that such questions would elicit responses which would provide information sufficiently accurate to meet users' requirements. More details of this process can be found in the 2001 Census White Paper (Cm 4253).

Cost of alternatives

I am not readily able to provide cost estimates of a sample census on the bases that you have suggested, nor would any such comparison with the costs of a full census be meaningful since a small sample would not facilitate comparable outputs. It would be the case, 107WA however, that even for a household sample of this size, the costs of certain activities, such as consulting with users on data needs, devising and testing questions, designing and implementing processing systems, publicity, and dissemination would not be significantly reduced; and that other costs covering the fieldwork and processing would not be reduced in proporation to the sample size. Additional costs would, indeed, be incurred in, for example, the requirement to design an appropriate sampling frame.

Advantages

The advantages of a full statutory census are that the data collected are comprehensive, authoritative, interrelated, consistent across different areas, and relate to a common timeframe and represent value for money.

The disadvantages might be perceived to be that it involves a relatively rigid process, that the balance between continuity and relevance is often a difficult one, that it is a very large investment, and that the results may not be immediately available.

Post-census evaluation

There are plans to conduct a post-census evaluation of the cost-effectiveness of the exercise in the light of the need to make recommendations on the viability of alternative means of collecting comparable information in the future, as well as on the form that any future census might take.