27 September 2010

This Statement sets out the Statistics Authority’s position in relation to numerical information that is not currently treated as official statistics.  Typically this will be described as management information or research findings.

The Authority acknowledges that government bodies need to collect management information to run their own businesses and will often wish to make it public in the interests of transparency.  Similarly, various kinds of quantified research findings are produced and published by a wide range of bodies.  Whilst both types of data are statistical in nature, they are often not regarded as ‘official statistics’ and, where this is the case, the Code of Practice for Official Statistics does not constrain their collection or publication.  In particular, two important requirements of the Code do not apply:

• the requirement to publish the statistics separately from political comment, and;

• the requirement to publish in the form of a statistical release before any political or other public use is made of the statistics. 

However, there are circumstances under which the Statistics Authority will regard some management information – particularly in aggregate form - as being official statistics; and to expect it to be produced in accordance with the Code.  In some fields this is common practice. Crime records, hospital records, school records, social security records, among others, begin, at the local level, as management information and are then aggregated to create some of the most high profile and important statistics.  This raises the question of which sets of aggregate management information should, in the public interest, be treated as official statistics, and which may reasonably continue to be produced without reference to the Code.

The National Statistician has issued guidance to departments on the principles to which government bodies should have regard in deciding whether or not a particular set of data should be treated as official statistics.[1] This guidance acknowledges that not all cases will be clear-cut; and there may be cases in which the Statistics Authority will conclude that data that have not previously been published as official statistics should in future be treated as such, and that, in the interests of maintaining public confidence in important statistics, the requirements of the Code should be observed.

In deciding whether to propose to Ministers that a set of data should in future be treated as official statistics, the Statistics Authority will place particular weight on two considerations:

• whether the data are used publicly by the department in support of major decisions on policy, resource allocation or other topics of public interest,  or;

• whether the data attract public controversy when published and the Authority takes the view that public debate would be better informed if the figures were, in future, handled as official statistics.
  

An exception would be made to the first point above where the Authority believed that, despite the public use, the data were of a kind that was unlikely to be thought of as official statistics by Parliament or the public.

Where the Authority concludes that the data should be published as official statistics, it will write to the Minister, or senior official, responsible for the producer body to explain its view. Depending on the circumstances, the Authority may further comment on the manner in which the figures have previously been published and indicate respects in which the requirements of the Code of Practice appear not to have been met. All such correspondence will be published by the Authority.

However, the decision on whether to publish data as official statistics remains a matter for the producer body. The Statistics Authority does not, except in the case of the Office for National Statistics, have statutory authority to direct public bodies on this matter.

This Statement will be kept under review and amended as necessary.

 

ANNEX - DEFINITIONS AND TECHNICAL CONSIDERATIONS

For the purposes of this Statement, we define management information to include all numerical data produced, or publicly used, by official bodies but which that body does not regard as official statistics. It thus includes the following:

• departmental business data that are not regarded as official statistics; 
• quantified research results (for example the results of a one-off survey) where the body responsible does not regard those results as being official statistics;
• raw statistical data that, whilst not themselves published as official statistics, are the raw material from which official statistics are produced (for example, local counts of hospital records). These are covered by some aspects of the Code of Practice and there are therefore some particular considerations discussed further below;
• special aggregations of raw statistical data, distinct from those aggregations that are published as official statistics (for example, figures relating to a particular area of the country when that area is not one of the standard areas used for the official statistics). Again there are particular considerations, discussed below.  

Raw statistical data

Raw statistical data will usually be the local or low level statistics from which official statistics - in their published form - are compiled. Here the detailed data, whilst not regarded as official statistics themselves, are nonetheless subject to some of the requirements of the Code of Practice – for example the assurance of impartiality, quality standards and protection of confidentiality. The Statistics Authority expects the related official statistics to be published before, or at the same time as, the more detailed breakdowns but this is not an absolute rule. There have, for example, been cases of raw data being obtained and released under the Freedom of Information Act before the related statistics are due for publication. While rights under FOI must be respected, the Authority will want to consider such cases and see what lessons might be drawn about good practice for the future.  
 
The Authority will investigate any release of, or public reference to, raw statistical data prior to the release of the related official statistics. It will then consider with the responsible body the particular circumstances of such cases and any lessons that may need to be drawn. 

Special aggregations

Bodies that produce official statistics sometimes issue special aggregations of raw statistical data, distinct from the official statistics that are normally produced from those data, in order to address a specific topic. The Statistics Authority will normally regard such special aggregations as being official statistics and will be unlikely to accept the argument that they are distinct from the normal official statistics and can therefore be released without reference to the Code of Practice. Should there be a case for treating special aggregations as other than official statistics, the Authority would expect the National Statistician to be consulted in advance of their publication.

---