Response submitted by MeCCSA to the Better use of data in government public consultation

Better use of data in governmentSubmitted 22nd April 2016

This response to the Better use of data in government public consultation is on behalf of MeCCSA (the Media, Communications and Cultural Studies Association) which is the subject association for academics and students in these fields in UK Higher Education.

You can read the full response below or download a PDF version.

We have focussed our response on the two final sections, questions 15-20, which are of direct concern to MeCCSA members.

The consultation document states that the “These proposals are not about selling public or personal data, collecting new data from citizens or weakening the Data Protection Act 1998” (p. 3). Whilst we welcome this stance, we remain concerned that once such a system of cross-referenced and linked data is established it opens the possibility of this position changing in future. In other words, safeguards must be put in place to prevent future commercialisation or undue and prejudicial protection of public or personal data, and erosion of the Data Protection Act, by a new Government.

The proposed “shift towards querying datasets through APIs in place of the typical practice of using bulk data shares” (p. 3) is welcome, provided there is flexibility within this to enable bulk data download should it be required for research purposes.

MeCCSA is concerned that the proposals determine that “data can only be disclosed if all the participating bodies and individuals are accredited” (point 102, p. 29) and alarmed that this extends to “the researchers, the accredited access facilities and the research itself” (point 103, p. 29). The language of the Impact Assessment pertaining to disclosure of de-identified data is vague and is already aligned towards specific disciplines in the way it describes even the most generic accreditation assumptions. A great deal of both quantitative and qualitative data on public behaviour, attitudes, or requirements in areas related to communications, culture, and media is held by statutory bodies (and it is not clear if quasi public bodies such as ofcom would be included in these proposals) and it would be enormously damaging to UK research to place any unnecessary restraint on the availability of such data.

Whilst we agree it is important that the accreditation body “should have expertise in statistical research and analysis” (point 104, p. 29), any organisation tasked with managing or preparing data should also give due consideration to qualitative research and non-statistical methods of analysis. Such data as that held in, for example, Family Spending, even the census data or in such aggregate summaries as Social Trends, are regularly and necessarily used by members. The more fundamental point is that whatever the expertise of an accrediting body, the proposed  procedure would be wholly incompatible with free academic debate, in which dispute about the valid, accurate, or proper  use of primary data is itself a matter for discussion in the public domain between researchers, not a matter for pre-judgement by an accrediting body, least of all one which is close to or part of the body compiling the data.  The expertise of the proposed body is not the issue we question but rather its capacity improperly to pre-decide who would have access to data and how it would be used.

Better use of data presupposes making available structured data for research, though it is important that this does not restrict access to data that remains unstructured. That is, researchers must continue to be allowed access to public documents where they can extract and structure data in ways that may have been unanticipated when the legislation or accreditation guidelines were drafted. Moreover, it is important that when data is made available it is clear through meta-data as to both what data is included as well as being explicit about what data is excluded.

Access to data which must be linked and de-identified using defined processes for research purposes

Question fifteen: Should fees be charged by public authorities for providing data for research purposes, and if so should there be a maximum fee permitted which is monitored by the UK Statistics Authority?

(X) No

In our view linked and de-identified data should be made accessible for research purposes without the imposition of a fee. Whilst the consultation documentation states that any such fee would be “charged on a costs recovery basis only” (pp. 17 and 30), we remain concerned that this would place a prohibitive burden upon researchers within our field. There is scant detail about what a “commercial rate” for cost recovery might entail (either by the public body or the third party Indexer), and it is likely any such fee would restrict access to all but large-scale funded projects (e.g. RCUK or Horizon 2020). There is no certainty that any such costs would be accepted in research tenders by such bodies, and research inevitably and frequently results in data requirements that cannot be anticipated at the design stage. Such a system would undermine humanities research in particular, where research is often carried out by individual academics supported by institutional or QR funding allocations, and could have seriously deleterious impacts on preliminary, unfunded, or pilot research. In the very unwelcome event that a fee is nevertheless introduced, we strongly advocate the introduction of a category for non-funded research that offers either a fee-waiver or a symbolic maximum fee for providing data for research purposes.

 

Question sixteen: To ensure a consistent approach towards departments accepting or declining requests for disclosing information for research projects, should the UK Statistics Authority as the accreditation body publish details of rejected applications and the reasons for their rejection?

(X) Yes

The Government should be increasing transparency of information it holds, as well as its operational procedures and decision making processes to make information that is in the “public interest” public by default. To this end we support the UKSA publishing details of rejected applications. It is important that reasons provided for rejection are clearly given and there is adequate opportunity for appeal against decisions made. We remain concerned that Government departments might use the cover of a permissive power to reject applications for research that might question current policy, or does not serve their political interests or priorities. Research ‘in the public interest’, even were this presumed to be a criterion of judgement, should not be equated to ‘aligned with current policy’.

 

Question seventeen: What principles or criteria do you think should be used to identify research that has the potential for public benefit, or research that will not be in the public benefit?

The contribution of research to the public good is now included as a criterion of judgement in the assessment of excellence in university research undertaken by the Research Excellence Framework.  It cannot and should not be pre-judged prior to the conduct of the research, and there is the very real risk that research by ‘accredited’ researchers, using data approved and released by a parti-pris accrediting body, would impede the credibility and authority of any research arising, and in turn the authority and reputation internationally of UK research. Research is either excellent or less so in its contribution to knowledge and understanding, and advance in research can only be impeded by prior selection of proposed research by holders of relevant data or their proxies. No accrediting body of the kind proposed can possibly have a credible and legitimate role in pre-judging the value of future research.

 

Question twenty: What principles and factors should be considered in preparing the Code of Practice on matters to be considered before making changes to the processes that collect, store, organise or retrieve data?

  • Sensitivity towards personally identifiable information
  • Stressing the importance of access to data – and, crucially, metadata about what the data covers AND does not cover. The latter is all too often missed, allowing Government to swamp open data repositories with useless data, giving the impression of transparency when actual meaningful data is withheld.
  • We support direct access to data for academics, to circumvent possibility of politically motivated vetoes

 

Posted by Einar Thorsen