Research Microdata Files (RMFs) are unit record files provided for statistical research purposes by the Central Statistics Office (CSO) under Section 20(c) of the Statistics Act, 1993. While RMFs do not contain direct identifiers, the risk of disclosure through indirect identification may be significant. The processes for authorising access to RMFs and for managing RMF research projects are therefore strictly controlled by the CSO.
RMFs are not statistical products. Unlike statistical products which relate to aggregated statistical analysis, RMFs are not published or made available to the general public.
The legal framework under which access to RMFs is granted is Section 20(c) of the Statistics Act, 1993. This section enables the Director General to appoint persons as Officers of Statistics
“to perform for a specified period particular statistical analysis which may necessitate access to data collected under this Act”.
Access is granted for research purposes, subject to a strict application of Section 20(c). This access must relate to a particular statistical analysis and it must be for a specified period. Thus every appointment of an Officer of Statistics under Section 20(c) is:
The researcher must complete a Declaration of Secrecy as set out in Section 21 of the Statistics Act before access to an RMF is granted. The researcher is obliged by law to respect the statistical confidentiality of information contained in the RMF (Section 33) and may only use that information for statistical purposes (Section 32). Any breach of these requirements is an offence under Section 38 of the Act and may be subject to prosecution.
Requests for access to RMFs must be made using the RMF application form. The ultimate discretion regarding the provision of access to RMFs rests with the Director General of the CSO.
The CSO emphasises that Officers of Statistics are legally obliged to ensure the confidentiality of RMF data. As part of this, persons applying for access to RMFs are required to demonstrate their knowledge of statistical disclosure control and to apply these methods to all tables intended for dissemination.
Any discussions of the data by the researcher (e.g. discussions of tables or analysis which could potentially disclose details of individual records) must be restricted to other Officers of Statistics appointed to the same statistical research project. The CSO also has the right to perform any appropriate statistical disclosure control, either before the RMF is issued to the researcher, or to any subsequent output generated from the RMF. This does not lessen the aforementioned obligations on the researcher appointed as an Officer of Statistics to perform all necessary statistical disclosure control. Failure to do so will result in CSO sanctions.
Access to RMFs is a privilege granted to researchers who meet conditions and criteria as designated by the CSO. This policy sets out the conditions under which such access may be granted.
The application for access will be assessed vis-à-vis the RMF assessment criteria. The research proposal must state in sufficient detail:
a) the purpose of the research
b) the explanation as to why this purpose cannot be fulfilled using non-confidential data
c) the entity requesting access
d) the individual researchers who will have access to the data
e) the individual researchers’ proficiency in the use of statistical disclosure control
f) the access facilities to be used
g) the datasets to be accessed
h) the methods of analysing the datasets
i) the intended results of the research to be published or otherwise disseminated.
Access, subject to adherence with the Statistics Act, 1993 and the associated policy and protocols for accessing RMFs, will, in general, only be granted to:
a) researchers who, either in their own right or as employees of a recognised research organisation/institution, have a proven track record in data analysis or research
b) researchers or individuals working in organisations/institutions that can give a clear and specific research rationale, acceptable to the Director General, for access being granted
Access to RMFs will not be granted to:
All access for researchers to RMFs, with the exception of Growing Up in Ireland, which is covered by the terms of the previous RMF Policy, will be controlled by the CSO by means of a Virtual Desktop Infrastructure (VDI). In the VDI, the analysis of RMFs takes place within the CSO ICT systems. Access to the VDI may be provided on-site in CSO or off-site, at the sole discretion of the CSO. Certain categories of RMF may only be accessed on-site.
The CSO VDI is a secure remote system through which approved researchers access microdata. There is no facility to copy data from the VDI to the local client PC.
Software applications provided for data analysis will be limited to a standard suite pre-installed on the virtual desktop.
For the purposes of this policy, off-site relates to premises other than the CSO offices in Skehard Road, Ardee Road and Swords. There are two different access types which are defined as follows:
a) On-site VDI: Researcher attends CSO office to access the RMF via VDI on a CSO PC
b) Off-site VDI: Secure remote access (via VDI) to RMFs residing in the CSO
Criteria for determining type of access include:
a) The security and suitability of the working environment of the researcher
b) The researcher’s affiliation with a recognised organisation/institution that has a proven track record in data analysis or research
c) The indirect disclosure risk of the data included in the RMF
A meeting with the CSO’s Data office staff will be required. The purpose of this meeting is to provide a comprehensive reinforcement of the RMF agreement conditions.
Access to RMFs for students will be restricted to those undertaking at a minimum post-graduate work and in all such cases their supervisor(s) must also apply and be appointed as an Officer of Statistics before access can be granted.
Access to the RMF, where granted, will only be provided once the researcher(s) has:
a) been appointed as an Officer of Statistics, and
b) signed the Declaration of Secrecy, and
c) formally agreed to abide by the RMF Standard Agreement, and
d) attended a training course provided by the CSO Data Office to provide a comprehensive reinforcement of the RMF agreement conditions which the researcher has accepted and to set out the legal obligations under the Statistics Act which a researcher undertakes when appointed as an Officer of Statistics.
The CSO will maintain a detailed register of individuals appointed as Officers of Statistics for the purpose of accessing RMFs.
Requests for the provision of RMFs to or from locations outside of the Republic of Ireland will, in general, not be facilitated.
RMFs sourced from administrative data will consist of sample data only and may be subject to additional terms and conditions as designated by the CSO.
In general, access to RMFs relating to business surveys and administrative data will be allowed only on-site.
IT Security and Access
In relation to VDI access:
a) Only those persons appointed as Officers of Statistics may use the remote access service.
b) CSO will request the IP address of the researcher/institute prior to account set-up and access may be restricted based on the specified domain which must be a main fixed business IP address.
c) The service must only be accessed from a secure location such as office or research facility.
d) The CSO reserves the right to audit the procedures in place at the off-site location, without prior notification, to ensure that the appropriate procedures are in place to protect the confidentiality and integrity of the data.
e) The researcher needs to ensure that access to the service is at all times restricted to the appointed Officer(s) of Statistics. This includes ensuring that the system is logged off when not in use and the data cannot be viewed by anyone other than the appointed officer.
f) Log on credentials must not be stored, shared or otherwise communicated.
g) The fob should be stored safely at all times and must be returned upon expiry of appointment.
h) Recording, copying or attempting to transfer data in any format from the VDI is strictly prohibited.
i) Any breach of the above must be reported immediately to the Researcher Coordination Unit (RCU) of the CSO.
On-site VDI access can only be permitted via a CSO desktop computer using the VDI infrastructure with all external media drivers disabled.
Responsibility for ensuring the confidentiality of all outputs (reports, publications, presentations, articles, etc.) based on the research carried out on the RMF (or using any element of the RMF) rests with the individual appointed as an Officer of Statistics. The restrictions and prohibitions on disclosure of information are set out in Sections 32 and 33 of the Statistics Act, 1993.
It will only be permissible for the researcher to take non-confidential aggregate data off-site for further analysis. The relevant Statistician/Senior Statistician will review aggregates from the VDI to check that they are non-confidential in nature. These must be assessed in the context of all available aggregate information to guard against disclosure through comparing different aggregates.
CSO staff will release the relevant aggregates to the researcher only after sanction to do so has been given by the relevant Statistician/Senior Statistician indicating that the aggregates checked are of a non-confidential nature.
All outputs from the research project should be provided to the CSO for information purposes. If required by the CSO (specified in the Standard Agreement), all outputs (reports, publications, presentations, articles, etc.) must be submitted to the CSO for approval, prior to being put into the public domain (i.e. communicated to anyone who is not an Officer of Statistics) so that adherence to the Statistics Act, 1993 and the protocols attached to the assignment can be assured.
The CSO reserves the right to put outputs from the research into the public domain if the researcher (individual appointed as an Officer of Statistics) has not already done so.
The RMF, regardless of any amendments made during analysis by the researcher(s) will at all times continue to be the property of the CSO.
The analysis/research undertaken must comply or be consistent with the specific purpose for which the access was granted.
Persons appointed as Officers of Statistics must not attempt to match/link (at a micro level) the RMF to any other non-CSO data source. Linkage to other CSO data sources is only permissible subject to the written agreement of the CSO. Any dataset derived from a CSO RMF by means of such linkage is subject to the same confidentiality requirements and conditions of use which apply to the original RMF.
On completion of the research or termination of the conditions relating to the appointment of the individual(s) as an Officer of Statistics, VDI access will be terminated.
In addition to the output safeguards described earlier, analysis sessions may be recorded by CSO staff to ensure that the RMF policy is fully adhered to.
The CSO reserves the right to monitor the implementation by researchers of their RMF agreement with the CSO, up to and including audits by the CSO, access logging and recording of VDI sessions.
The CSO must be acknowledged as the data source in all outputs including citing in footnotes to aggregate tables and analysis based on derived variables. The following citation must be included in all publications:
“Results are based on analysis of strictly controlled Research Microdata Files provided by the Central Statistics Office (CSO). The CSO does not take any responsibility for the views expressed or the outputs generated from this research.”
Failure to comply with the protocols, terms and conditions specified in the standard agreement attached to the provision of RMFs may have implications for the individual and the organisation/institute for whom they work. These sanctions may include but are not limited to:
a) Termination of the individual’s appointment as an Officer of Statistics
b) Requirement to return and/or cease using all information provided by the CSO
c) Corresponding sanctions in relation to the organisation/institute and other RMF researchers in that organisation/institute
d) Denial of future requests for RMF research access
The CSO reserves the right to apply other sanctions, up to and including prosecution under the Statistics Act, 1993, where appropriate.
Where it is practicable and feasible, data requests will be met via the provision of aggregate data (such as anonymised microdata files (AMFs)) as distinct from the provision of a RMF.
The CSO will apply a principle of data minimisation when providing access to RMFs. I.e. the information provided to the researcher will be limited to those topics/variables in the survey which are necessary to the specific statistical research project.
The policy for business statistics RMFs is that the unique identifier is pseudonymised.
