Memorandum by the Demographics User Group
The Demographics User Group represents fourteen
large commercial users of Census and other government data. Its
membership comprises: Abbey National, Bank of Scotland, Bradford
& Bingley, Halifax, Marks & Spencer, Marks & Spencer
Financial Services, Nationwide, One 2 One, Royal Bank of Scotland,
Saga Group, Sainsbury, Tesco, Whitbread and Yell.
DUG welcomes the opportunity to submit evidence
to the Committee under the four headings:
1. THE ADEQUACY
The Office for National Statistics is to be
congratulated in the efforts that they have made to maximise the
value of the Census to users. As with previous Censuses, there
were strong contacts with established users in central and local
government and the universities. For 2001, successful efforts
were also made to increase the involvement of commercial users
of the Census.
Our only serious regret was that, despite the
wishes of all the major user groups, the ONS appeared to have
decided at the outset not to ask a question on Income.
2. THE CONDUCT
The major fear of users is always that any press
campaign against the Census might cause widespread refusal to
respond, resulting in inadequate statistics. Thankfully this did
not occur, although we did feel that the ONS might have taken
a more proactive stance in explaining the benefits to the press
during the key Census week.
We were also nervous that the new procedure
for forms to be posted back, rather than collected, might lead
to reduced response. These fears have proven to be unfounded,
and we understand that response is c.98 per cent, as in 1991.
Anecdotal evidence raises some doubts about
the completeness and quality of responsesthe usual Quality
Survey should have been held after the Census.
However, for most commercial organisations which
use the Census to support business decisions, such defects are
vastly outweighed by the value of the data that has been collected.
Several innovations will be of great value to
OUTPUTSTHE MAJOR ISSUES FOR USERS
|Issue ||DUG's view
|100 per cent Coding and|
|The move away from 10 per cent sampling is a great improvement for many users who analyse statistics for small areas.|
The approximation to Social Grade is very welcome.
|Geography||Another major advance, enabling postcoded customer data to be related to small Census areas, and digital boundaries and background map data to be available from the ONS
|Free dissemination of statistics promises to be a truly revolutionary advance, and will greatly increase use of the data.|
This will be reinforced by inclusion of the Census within the wider framework of Neighbourhood Statistics, and the sharing of the common Output Area geography.
|ONS's effort put into consultation has been appreciated (although it would have been less time consuming to ask users of 1991 statistics their views, rather than invent new tables in a back room and then seek comment).|
However, one serious concern that has emerged in the last few weeks is that ONS's recent preoccupation with Disclosure Control might be inventing theoretical problems that have never arisen in decades of practical experience. Excessive Disclosure Control measures could seriously reduce the value of Census data, particularly for small areas.
One particular instance is the apparent reluctance to produce statistics aggregated to postal sector level, which was done with no adverse consequences in 1991.
It is significant that the Scottish Census Office are planning to continue tried and tested approaches used in previous censuses.
News of the ONS's performance in processing the data and
delivering statistics is currently awaited. The speed of delivery
appears to be slower than in other countries, but it must be recognised
that any processing errors and re-issues (as happened in 1991)
cause huge disruption for users.
4. LESSONS FOR
One particular lesson to learn for the next Census is that
the coverage of empty property/second homes/visitors in 2001 seems
to have been erratic.
More significantly, the cost of the Census is very modest
in comparison with the magnitude of the financial decisions (by
both public services and commercial companies) which are based
on its counts, and the case for holding the next Census in 2006,
rather than 2011, should be assessed without delay.
9 November 2001