APPENDIX 3
Memorandum from the Royal Statistical
Society
1. INTRODUCTION
1.1 The Royal Statistical Society has, since
1990, argued that a national statistical service, free from political
interference, is vital to the maintenance of a healthy democracy.
The appointment of a National Statistician, the appointment of
a Statistics Commission, and finally the publication of the Framework
for National Statistics and the accompanying Initial Scope
of National Statistics in June 2000 are the first steps towards
establishing such a national statistical service. However, there
is much work still to be done to provide government, Parliament,
researchers, the general public, and other users of statistics
with a comprehensive range of statistics about the UK which are
of an appropriate quality and fit for purpose.
1.2 The Royal Statistical Society takes
heart that the Framework for National Statistics document
is a "first edition", and hopes for further editions,
clarifying and improving the framework for National Statistics.
Similarly, the title of its accompanying document—Initial
Scope of National Statistics—is absolutely crucial.
The present scope of National Statistics should be considered
only as a starting point for the much-needed further expansion
of the arrangements.
1.3 This evidence is therefore shaped by
the Royal Statistical Society's definition of National Statistics,
as set forth in its response to the 1998 Green Paper Statistics:
A Matter of Trust:
"The scope of National Statistics should
be inclusive of all statistics of public interest at the national,
regional and local levels, whether they are produced by government
departments, agencies, regional or local governing bodies, or
any other public sector body. The needs of users, as well as producers,
should inform the scope of National Statistics."
1.4 The general rule for inclusion within
the Initial Scope has been one of existing Government Statistical
Service (GSS) products. This is acceptable as a starting point,
but it does not support the concept of National Statistics as
a service which provides the full range of data which allow the
citizen to judge the state of the nation, as there are too many
high profile omissions.
2. SCOPE OF
THE NEW
ARRANGEMENTS
2.1 What is a National Statistic?
The Initial Scope of National Statistics
is defined primarily in terms of publications. Publications
may have aspects of their content changed, but still remain the
same publication. If the content of a database is changed, then
it is no longer the same database. If publications, rather than
databases, are included, will it be possible to alter the content
of publications while keeping them within National Statistics?
There are no clear rules. The Royal Statistical Society would
argue for a more holistic approach which relates to data systems
rather than publications, particularly as publications policy
is likely to change. The Department for Education and Employment
has included both databases and publications within National Statistics,
and this is something which we would support.
2.2 Management data and statistics
One area of ambiguity that hampers a clear definition
of National Statistics is the vague borderline between management
data and statistics. For example, it might be argued that monthly
hospital waiting list figures are data which allow managers to
audit and improve the performance of various systems. On the other
hand, members of the general public, the media, and other interested
parties might argue that they would use monthly waiting list figures
to judge the effectiveness of government measures to reduce hospital
waiting lists. Given the fact that they are the subject of high
profile media releases, there might be no doubt in the mind of
this latter group that monthly hospital waiting list figures are
used to determine the state of the nation and are, therefore "National
Statistics", yet they have been excluded from the initial
scope.
2.3 Public use of National Statistics
It is not unreasonable to define National Statistics
as data which are used by stakeholders, including the public,
to judge the state of the nation and the performance of government.
In that case, another important topic to consider would be how
should public use be defined?
2.4 It is, of course, beyond the power of
the producer to limit the uses to which statistics are put. For
example, league tables may be used by politicians, the media and
members of the public, despite any caveats placed upon them by
the producer of the underlying data, because they appear to aid
the decision making process. League tables[2]
are excluded from National Statistics—rightly so at the
present time, given that they can be very misleading and, as they
are based on small numbers of data items, they can exclude a wealth
of additional information.
2.5 One of the aims of National Statistics
listed in the Framework Document is:
"To inform . . . the citizen about the
state of the nation and provide a window on the work and performance
of government allowing the impact of government policies and actions
to be assessed". This is a positive step away from the ethos
that official statistics are produced for the use of government
alone. However, if apparently useful datasets are excluded from
National Statistics, what is the citizen to make of the role and
use of National Statistics? The Royal Statistical Society would
advocate a systematic examination of statistics against clear
criteria to see which should, in principle, be included as part
of National Statistics, subject to matters such as quality. This
would hopefully lead to the elimination of inconsistencies where
high quality data from some major surveys has been excluded. The
underlying concepts behind National Statistics could be strengthened
by a full and open debate about how and why the public uses statistics,
perhaps leading to greater resources being devoted to bringing
statistics currently excluded on the grounds of poor quality,
such as league tables, up to the standard required for National
Statistics.
2.6 Government targets
The Government is keen to publish targets against
which the effectiveness of its policy initiatives may be assessed.
For one high profile target—to reduce class sizes—the
data to monitor it are included within National Statistics. Will
there be an undertaking from the Government that it will automatically
place data which will be used to measure all its targets within
National Statistics?
3. OMISSIONS
FROM NATIONAL
STATISTICS
3.1 Retail Price Index (RPI)
The Royal Statistical Society reiterates its
dismay that the Chancellor of the Exchequer, the Minister for
National Statistics, has kept the RPI out of National Statistics.
The Economic Secretary to the Treasury herself described the RPI
as "of special importance to the UK economy"[3].
The exclusion of the RPI weakens the image and standing of National
Statistics as a service which informs the Parliaments and Assemblies
and the citizen about the state of the nation.
3.2 Surveys
There is considerable inconsistency over inclusion
of surveys. Outputs from the programme of National Surveys of
NHS Patients, funded by the Department of Health, are not included
within the Initial Scope, whereas the Health Survey for England
is. It is not clear why one survey has been included and the other
not, when they seem to have the same characteristics—both
being large scale, and funded by the Department of Health on a
regular basis.
3.3 Other surveys done regularly but not
annually, such as the five-yearly Infant Feeding Survey[4]
and the 10 yearly Survey of Adults and Children's Health, have
been excluded.
3.4 The consistent line across all departments
has been that only regular time series were included in the Initial
Scope. No department has been brave enough to make a statement
in the Initial Scope to the effect that any future survey or other
statistical work undertaken to research an issue of national concern,
even if it is a one-off project, will be included within National
Statistics. The Royal Statistical Society would like to seek assurances
that decisions about the inclusion of survey results in National
Statistics will be made at the time of commissioning each survey.
3.5 This patchy coverage of National Statistics
does not add weight to the Government's commitment to evidence-based
policy. Statistics about new policy initiatives will be published.
For those statistics to be outside the scope of National Statistics
will undermine both the standing of National Statistics and the
Government's claim to have removed the potential for political
interference.
3.6 Monthly Hospital Waiting List Figures
Quarterly hospital waiting list figures are
included in the Initial Scope, while monthly waiting list figures
are not. Monthly waiting list figures have been classified, it
would seem, as management data. A possible reason for the exclusion
of the monthly figures would be concern that they contain too
much measurement error, and that they have too high a political
profile, to allow them to be included within National Statistics.
The very fact that they have a high profile lends support to the
need to ensure that they are of a good standard.
3.7 The National Statistics code of practice
will include guidance on release arrangements. Statistics should
be released at a pre-announced time, which is not varied for political
advantage. Ministers should not comment until the statistics are
released and in the public domain, creating a level playing field
for all users. If, in two months out of three, these practices
do not apply, and in particular ministers leak the waiting list
statistics or comment in advance of the release, then public confidence
in National Statistics will be undermined. The public will not
distinguish the one month in three when the release is a National
Statistic from the two months in three when it is not. Poor release
practices in this case will undermine confidence in National Statistics.
3.8 Statistics relating to areas of public
concern
The first edition of the Framework Document
has appeared to define National Statistics as the outputs of the
Government Statistical Service (GSS) This is far narrower than
"all statistics of public interest". There are many
areas of public concern which fall outside the remit of the GSS.
The Royal Statistical Society believes that the future scope of
National Statistics should not be restricted by perceptions of
what work the GSS performs, or has performed in the past.
3.9 The National Statistician should have
responsibility for ensuring that the necessary statistical expertise
is recruited and properly involved in areas which arguably arouse
the greatest amount of public concern, debate and fear, such as
BSE/vCJD, MMR vaccination, GM crops and food, and passive smoking
and health, where the predominant problem is scientific uncertainty.
The conclusions of the modelling and data analysis must be made
public in a rigorous but digestible form.
3.10 Statistics across the UK which cannot
be aggregated
There are a number of instances where there
is a commonly held belief that a particular statistic should exist,
and be regarded as a "National Statistic" yet, under
the present arrangements, it is impossible to produce that statistic.
One example is the Caesarean section rate for the UK, which cannot
be produced because data of sufficient quality are not available
for all countries, and across the countries there are differing
bases for publication. This makes it impossible to combine figures
from England, Wales Scotland and Northern Ireland to obtain a
Caesarean section rate for the UK. This was the situation even
prior to devolution.
3.11 The Royal Statistical Society believes
that for important figures it ought to be possible to produce
a UK figure, and individual country data should be made available
in such detail as will enable the production of the UK figure,
by outside analysts, if not the GSS.
4. AREAS REQUIRING
CLARIFICATION
4.1 Accessibility of data
Many services which were once delivered by the
public sector are now delivered by the private sector, or by agencies
instead of government departments. The National Statistician ought
to be able to demand statistical information from any organisation,
just as (s)he does from any company under the Statistics for Trade
Act, to compile National Statistics for publication. This should
apply to a public sector organisation, for example, a hospital
or a school, or a quango or quasi public organisation, for example,
the Civil Aviation Authority, the BBC, the Health and Safety Executive,
just as it does to a company.
4.2 This would correct the major anomaly
which is the omission of Scottish health statistics from National
Statistics. Scottish health statistics are produced by the Information
and Statistics Division of the Commons Services Agency for Scotland.
This is part of the NHS, rather than the GSS. Scottish health
statistics should not remain excluded from National Statistics
on the basis that the producers are part of the NHS and not the
GSS.
4.3 An organisation which produces a set
of statistics should not automatically be the vehicle for publishing
National Statistics, although it could be if the National Statistician
were satisfied about the quality control processes and practices
within the organisation.
4.4 Financial considerations
The fundamental feature of the White Paper Building
Trust in Statistics and the subsequent new arrangements is
a focus on statistical outputs. Fortunately, in the area of quality,
it is already recognised that it is the whole statistical process
that matters. The Royal Statistical Society would argue for additional
funds to be made available to the National Statistician to improve
data collection processes in government departments, agencies
and other public sector bodies and thereby improve the quality
of their statistical outputs. One desirable outcome would be that
outputs currently excluded on grounds of quality would be brought
up to a standard which would admit them to National Statistics.
4.5 The Framework Document states that "Ministers
will take decisions about the coverage of National Statistics
in the light of the costs and benefits involved". Does this
mean costs and benefits to the department, or would benefits be
more broadly defined? This appears to give Ministers the opportunity
to, in a worst case scenario, remove a statistic from National
Statistics in order to ease pressure on a departmental budget.
4.6 One of the responsibilities of the Statistics
Commission listed in the Framework Document is to "consider
and comment to Ministers on the high level programme for National
Statistics, drawing on the views of users and suppliers, taking
account of: (i) the resources available for National Statistics
. . ." This implies that the Statistics Commission must regard
the level of funding available for National Statistics as an inflexible
amount of resource. This is far from giving the Statistics Commission
powers to comment upon the level of funding for National Statistics
and the right to argue for increased funding, or additional monies
for a specific project of national importance.
4.7 The Framework Document states that the
National Statistician has "the right of access to the Prime
Minister, through the Head of the Home Civil Service, on matters
concerning the integrity and validity of official statistics including
on resources where he/she believes they impact on the integrity
and validity of official statistics". If "official statistics"
are interpreted as the present scope of National Statistics, then
this does not give the National Statistician the right to request
more funds to extend the scope of National Statistics. Neither
the Commission nor the National Statistician have the power to
argue for increased funding. The Royal Statistical Society is
concerned that the scope of National Statistics could remain static,
or, in a worst case scenario, be reduced, as a result of funding
constraints.
4.8 Possible criteria for inclusion in National
Statistics
In paragraph 2.5 of this evidence, the Royal
Statistical Society has argued for a systematic examination of
statistics against clear criteria to see which should, in principle,
be included as part of National Statistics, subject to matters
such as quality, for which there should be separate criteria.
As a starting point for discussion, the Society would like to
suggest the following:
Would stakeholders, including the general public,
use the statistics to judge the performance of government?
Would stakeholders, including the general public,
use the statistics to judge the state of the nation?
Would the statistics be used to formulate or
monitor policy?
Is a particular statistic used by the political
parties to support or repudiate a claim of a successful policy
initiative?
Is the statistic a performance measure of a
particular policy initiative?
Do the statistics relate to matters which arouse
great concern, debate or fear among the general public, and where
a clear statement of the scientific uncertainty relating to the
matter under investigation would greatly help public understanding?
Is it reasonable to expect an easily aggregated
UK statistic to be compiled from the individual countries' statistics?
What should not be a criteria for inclusion
within National Statistics is whether the issue or subject under
investigation is within the public sector.
5. ABSENCE OF
A CODE
OF PRACTICE
5.1 The Framework Document made the National
Statistician responsible for preparing the Code of Practice. Given
that the National Statistician was able to take up the post only
a short time prior to the launch of National Statistics, this
has had the unfortunate side effect of those responsible for the
production of National Statistics having to work within the new
Framework, but without a code of practice. The Royal Statistical
Society feels that sufficient time has passed now, and would welcome
a statement about the timetable for the production for the Code
of Practice. The Society would expect to be consulted about its
content at the earliest opportunity.
6. ABSENCE OF
NEWS FROM
OR ABOUT
THE STATISTICS
COMMISSION
6.1 At the time of writing, the page about
the Statistics Commission on the National Statistics web site
had last been updated on 16 June 2000. There was no link to a
web site for the Commission. The Statistics Commission has a single
page, interim website dated 21 October 2000. There is an e-mail
address so that members of the public may contact the Commission.
However, there has been a long wait between the launch of the
arrangements for National Statistics on 7 June and the services
being put in place which allow the general public to contact the
Commission. This is not the fault of the Statistics Commission.
It was unfortunate that announcements about the Commission were
made before it had time to organise even its most basic requirements,
such as finding accommodation.
6.2 One of the responsibilities of the Statistics
Commission is to "ensure that it is able to assess the needs
of users". There has not yet been any announcement about
how the Commission intends to consult and involve users.
6.3 The Statistics Commission has no existence
in law, and therefore, in theory, could be stood down at any time
by the Minister for National Statistics. A practical problem for
the Commission which comes as a direct result of this apparent
impermanence, is the recruitment of staff to the secretariat.
The Chief Executive, for example, is on secondment from the Department
of Health. While the Royal Statistical Society does not question
the suitability of the current Chief Executive, it would be a
very great shame if, in the future, the important work of the
Statistics Commission was hampered by problems recruiting suitably
qualified staff.
7. ABSENCE OF
STATISTICAL LEGISLATION
7.1 The Framework Document First Edition
(4.2.5 [h]) states that the Commission should review the need
for statistical legislation after two years and report back to
the Minister for National Statistics. While accepting the pressures
on parliamentary time, the Royal Statistical Society strongly
believes, and has done so for many years, that legislation is
necessary to underpin the principles and practice of National
Statistics. Two years is too long to wait for a review of the
need for legislation. There is no statement of when a report on
the review would be published, and no indication of the scope
of the review. For example, the Royal Statistical Society hopes
that the review would include international comparisons. Much
publicity has been given to the new arrangements, yet they have
been left unprotected by law, and consequently, open to future
alteration.
23 October 2000
2 It is also beyond the power of the producer to prevent
a particular dataset being labelled a "league table"
by users. The Royal Statistical Society recognises this, and uses
the term to describe those datasets which are based on small numbers
of data items, and which exclude additional information. Back
3
Hansard Volume 336, No 133. Back
4
Formerly conducted by the GSS, the 2000 survey has been contracted
out to a market research company. Back
| 
