perspective in the release of micro data for scientific ...perspective in the release of micro data...
Post on 21-Apr-2020
2 Views
Preview:
TRANSCRIPT
Perspective in the release of micro data for scientific purpose
The new EU Regulation on access to confidential data for scientific purposes
Jean-Marc Museux (jean-marc.museux@ec.europa.eu)Aleksandra Bujnowska (alesksandra.bujnowska@ec.europa.eu)
Eurostat – Unit B2 methodology and research
LFS - EUSILC user conference - Mannheim - March 2011
Eurostat – Unit B2 methodology and researchStatistical Confidentiality and micro data release branch
Outline
� Background
� Principles of the revision of Regulation 831/2002
�
LFS - EUSILC user conference - Mannheim - March 2011
� Perspective opened by the revision of Regulation 831/2002
� Discussion - feedback
Background – EU statistics
� Harmonising MSs data collection establishing minimum common standards for EU statistics
� MSs collect data from data provider (individual-household, establishments)
� Eurostat gathering (micro) data from MSs and compiling and
LFS - EUSILC user conference - Mannheim - March 2011
� Eurostat gathering (micro) data from MSs and compiling and disseminating European indicators
� No further redistribution of raw (confidential) data except for scientific purpose but MSs should agree explicitly on each research proposal
� Provision of micro data for research purpose included in the core mission of the ESS– Maximise use of existing data – return on public investment– Value added of scientific research for policy analysis and for
quality of data
Background – Statistical confidentiality
� Ensure the right for confidentiality of information provider (individual, household – establishment)
� Impact on trust of information provider, non response and quality of information collected
� Risk of unlawful disclosure: reputation, business
LFS - EUSILC user conference - Mannheim - March 2011
� Risk of unlawful disclosure: reputation, business continuity, major loss for the EU
� Principles beard in European Regulations: – Regulation 223/2009 on European Statistics– Regulation 831/2001 defining rules and conditions for
access for scientific purpose
Background – ESS governance of statistical confidentiality
� European Statistical System Committee (ESSC –Chief Statistician in NSIs)
� Working Group on Statistical Confidentiality (WGSC
LFS - EUSILC user conference - Mannheim - March 2011
- MS delegates)� TF advising Eurostat for the drafting of the new
regulation – BE, BG, DE, FR, IT, LV, LU, HU, NL, SI, SE, UK
� Role of European Statistics Advisory Committee
The revision of regulation 831/2022 : the process
� 2010: Definition of the principles and strategic orientations of the new Regulation – TF, WGSC
� Feb 2011: Approval of the strategic orientations by ESS Committee – identification of remaining barriers
LFS - EUSILC user conference - Mannheim - March 2011
Committee – identification of remaining barriers � 2011: Drafting the legal text and accompanying procedures –
consultations� 2012: Adoption and entry into force
The revision of regulation 831/2001 : research community consultation
� Draw on a tradition of cooperation with research community (workshops, conferences, research projects)
� Advised by ESAC (J Leshke, L. Kabat, D Livesey)� Close link with the future Data without Boundary FP7
infrastructure research project (CESSDA network of data
LFS - EUSILC user conference - Mannheim - March 2011
infrastructure research project (CESSDA network of data archives)
� Why important at this stage– Trade off will have to be found to obtain a consensus among MS– Importance to be back up by Scientific Community for some
strategic choices
� Warning : no commitment possible at this stage of the process
The revision of regulation 831/2002 : principles
� Regulation is an enabling factor : should allow stepwise development of access in the next decade
� Regulation aims to provide minimum requirements regarding data access (what, how, to whom)
LFS - EUSILC user conference - Mannheim - March 2011
access (what, how, to whom)
� More transparency of procedures
� Objective: widen access to confidential data for scientific purpose– Through a multi mode approach– Involving different partners (DA, NSIs …)– Simplifiying rules, reducig administrative burden where possible
Strategic orientations of the new Regulation
� Introduction of the risk management approach for the release of confidential data: – safe people, safe settings and safe data principle, – Proportionality of control procedures
� Opening new modes of access to micro data, mainly
LFS - EUSILC user conference - Mannheim - March 2011
� Opening new modes of access to micro data, mainly remote access,
� Fostering of the decentralised access to confidential data
� Fostering partnership/delegation of tasks in the provision of access
� Streamlining the procedures for the release of low confidentiality impact level datasets (anonymised files)
Safe data principle : current
� Anonymised data (risk avoidance) release on CD-ROM under contract by Eurostat ( at risk of saturation)
–From 1.04.2011 Eurostat decided not to charge anymore the
LFS - EUSILC user conference - Mannheim - March 2011
–From 1.04.2011 Eurostat decided not to charge anymore the release of CD-ROM–Limited capacity to grow
� Detailed (raw) micro data available in Luxembourg safe centre – very few access requests
Safe data principle : target
� Datasets categorised into three categories taking into account – reidentification risk – perceived impact on the individuals concerned (sensitivity) – perceived impact on the ESS in case of unlawful disclosure of
confidential data
LFS - EUSILC user conference - Mannheim - March 2011
� The 3 confidentiality impact levels (CIL):
1. Low (less sensitive, anonymised files through information coarsening on key variables, outliers/population unique masking or perturbation)
2. Medium (more sensitive data, mode details in key variables) 3. High (raw data received from MS – no direct identifiers – full
information)
Safe data principle
� Procedures to be put in place should be proportionate to the risk and impact
� Low confidentiality impact level datasets released under simplified and timely procedures at zero cost
LFS - EUSILC user conference - Mannheim - March 2011
simplified and timely procedures at zero cost
� Besides confidential data channels : – Public Use Files (early 2012) – Online access (tables) (feasibility and pilot in 2011)– possibly remote execution (submission of batch)
Safe people principle : current
� European universities and research centres
� Lenghtly ( > 12 months) legislative procedure for accreditation other bodies (outside EU universities, Commission DG, …) –
LFS - EUSILC user conference - Mannheim - March 2011
other bodies (outside EU universities, Commission DG, …) –list published in the official journal
� Contract with research institution for each research proposal naming researchers
Safe people principle : target
� More focus on research proposal and principal researchers– Use and need for micro data– Publication provision and public benefit– Experience and professionalism of principal researcher– SDC literacy of principal researcher
LFS - EUSILC user conference - Mannheim - March 2011
� Institutional context will be integrated in screening procedure as a factor of risk – level will depend on type of access/data type – mission and organisation and purpose of research activities– measure in place for physical protection of data (only in case of
transmission of data)
Safe people principle
� Contract/license remains the basis for accountability of the researcher and for eventual sanctions (penal and administrative)
– Simplified procedure for low confidentiality impact level transmission :
LFS - EUSILC user conference - Mannheim - March 2011
transmission : • Framework contract • Automation of consultation of MS• Bilateral agreements with MS to wave out consultation on
specific datasets
Safe settings principle: current
� Transmission (off site) of anonymised files to admissible institutions
� Access in Eurostat safe centre (on site)
LFS - EUSILC user conference - Mannheim - March 2011
� Access in Eurostat safe centre (on site)
Safe settings principle : target
� Various modes of access – On-site – CIL3 – Remote access – CIL 2 ..– Off-site – CIL 1 ..
LFS - EUSILC user conference - Mannheim - March 2011
� Decentralised infrastructure – On site : accreditation of NSI safe centres– Accreditation of data archives and other partners
� Accreditation of access facilities agreed by the ESSC on a case by case basis according to criteria specified beforehand.
Main barriers
� Accreditation of non ESS bodies� Articulation with national law� The concept and implementation of confidentiality
impact level to be operationalised
LFS - EUSILC user conference - Mannheim - March 2011
impact level to be operationalised� Need to keep organisation signing the contract� Defining procedures and criteria for implementing the
regulation
Next steps
� Consultations with NSIs
� Discussion at the meeting of the WGSC in June 2011
LFS - EUSILC user conference - Mannheim - March 2011
� Discussion at the meeting of the ESSC in September 2011
� Voting: November 2011
top related